This paper discusses an improved colliding bodies optimization algorithm to pledge competently to realize optimal reactive power dispatch problem. In large scale power system, optimal reactive power dispatch problem is a huge constraint as it deals with large range of non-linear and non-convex global optimization problems involving the composite of continuous & discontinuous control variables. The reactive power dispatch problem is formulated by reactive power supply components like generator output voltages, regulating transformers and reactive power compensating devices. In order to get the optimal values of control variables, a robust optimization algorithm has been tested for reactive power dispatch problem. The prime objectives of reactive power dispatch problems are: to lessen the real power loss in transmission lines, reduce the congestion in transmission lines, to maintain the specified voltage magnitude in all the buses in power system network for both normal & abnormal operating conditions. This is achieved by satisfying the set of specified operational constraints. This algorithm was well-established on single-dimensional collisions intervening bodies, with every operator result being evaluated by substance with mass. Later, the collision of impelling bodies with titled masses & velocities showed that these objects were distinct with new impetus. This collision response prompts the operators to act in almost fine point in the search area. The optimization of the colliding bodies forward simple formations to catch minimum or maximum of functions. The algorithm was applied on standard IEEE - 57 bus system & the outcome solutions are correlated with lately matured algorithms like cuckoo search, particle swarm optimization and gravitational search algorithms. © 2017 IEEE.