Honda is one of the most popular algorithm for rear end collision avoidance system, but does not intend to avoid all the accidents, this due to that it gives a smaller range of warning which cannot avoid the accidents at high speed, to overcome this issue we proposing the optimisation of warning range with a genetic algorithm to minimise the probability of collision. The results show that proposed algorithm avoids 7% more accident than the standard Honda algorithm. We presented the MATLAB implementation of proposed work with examples to show the efficiency of our work.