can anybody share with me the best way of controlling exotherm in a batch reactor using cooling water as a cooling medium?

If the reaction is mildly exothermic, consider a pumped-loop water system, where the cooling water is recycled round the jacket or coils (etc.), with fresh water being added under reactor temperature control. So, if the reactor temperature has to be (say) 30 degC, you might find that the cooling water returning to the jacket is at 25 degC to get sufficient cooling.
Avoid accumulation of reactant(s) by some form of monitoring system e.g. thermal output.
If the reaction is very exothermic, consider operating at as high a temperature as possible, to absolutely avoid the possibility of accumulation which could then be followed by runaway reaction.
In general for this kind of thing you may need specialist advice including reaction calorimetry, to also look for the possibility of undesirable side-reactions or decomps.
Jill Wilday and co-workers at HSL Buxton produced a very nice Reaction Hazards workbook. (She's on LinkedIn somewhere)
You can also get advice from commercial labs.

You need to ask if the cooling water were to fail, would the exotherm merely give a spoiled batch or would it run away to an explosion. If the former, normal controls will do. If the latter you must take extra precautions and even consider your process. Do you dump everything in at the beginning, or can you add one component gradually while monitoring the temperature?
A cooling water jacket also depends for its effect on motion within the vessel, so failure of an agitator can also cause loss of cooling, perhaps disastrously.

The most effective option is to manage the reactor pressure at the boiling point and reduce the pressure thereby causing the reactor to boil. The heat transfer coefficent in a condenser is much higher than batch reactor. The reactor temparture control system needs to measure the reactor temperature, calculate the reactor pressure corresponding to that temperature and then adjust the pressure to cause the reactor to boil, condensate is then returned to the reactor cooling the reactor more effiicently then conventional jacket or externaly liquid/liquid heat exchange. PeterR

This technology was used very effectively to enable conventional polymerisation reactors to operate upto to 10 times faster than jacket cooled systems

it is the best way to control exothermic reaction by using shall and tube heat exchanger having water(cooling tower water) in shall and product you want to cool in tube to get best result.controlling will be so efficient by using instrumentation( controller and final controller element) you can say ,by using the close loop system.

Jacket cooled systems are often poorly efficient. Sometimes an efficient way to cool a batch reactor is pumping out the liquid, cooling it in an external heat exchanger (with cooling water available) and sending it back to the reactor. This possibility (and any other proposed here) can be advantageously modeled (in a dedicated batch chemical reactor simulation software such as BatchReactor from ProSim) prior to take a decision and implement it on your plant.

Jill Wilday will certainly point you in the right direction, The old ICI PVC reactors were water cooled jackets, and quite effective, others polymerisation reactors had pumped cooling systems. they certainly couldn't cope with a a cooling water failure and suitable relief is required, be aware of the possible need for multiphase flow relief modelling.
Standing on the reactor platform when the cooling water fails to 10+ PVC reactors is an interesting experience. The classic champagne effect is pretty violent!
Reaction kill systems can help protect against uncontrolled releases.
Vapours generated during the reaction process can be condensed and returned too.

Having had the reactor contents disappear up the column on one occasion, this is not a subject to be taken lightly. You need to know the dynamics of the system and all the eventualities before you commit yourself. In my case it was necessary to heat the contents with steam to trigger the reaction then put the cooling water on to control the exotherm. What we didn't know at the time was that the reaction rate was very sensitive to one particular reactant concentration. Fortunately no major harm was done.

John an additional feature in the control system was to control the boiling rate. An orifice plate in the vapour line would measure the vapour follow rate to the condenser. The reactor temperature control system would initially measure the reactor pressure and calculate the corresponding boiling point temperature and compare it with the actual temperature and temperature set point. If the reactor needed to be cooled it would firstly open cooling water valves to the reactor jacket and then vent the reactor to cause the reactor to boil. The boiling rate would be measured via vapour flow through the vapour line orifice plate. At full cooling the vapour rate would be limited to prevent boil over of the reactor contents. There are two sets of control valve used to control the reactor pressure either adding nitrogen to increase pressure or venting to reduce pressure. The PT control valves are larger than the control valves used to control the rate of boiling. The control system would use the larger valves to cause the reactor to boil and then switch to the smaller valves to control the boiling rate.

I am assuming the equipment is already existing and you are looking at safe and cost effective methods to better control the exotherm. Then might I suggest fitting a PID loop to the cooling water loop and increasing the agitation speed. This will give you closer control on your CTW loop and marginally increase the overall heat transfer coefficient, though the problem with higher agitation speed means the possibility of higher reaction rate. The other option is to scale down the process i.e. a smaller reactor has a larger surface area to volume ratio which has the effect of increasing the heat transfer capability of the reactor. As Martin points out, the rate of addition v's temperature kick could be examined.