In Particle Swarm Optimization (PSO) algorithm, although taking an active role to guide particles moving toward optimal solution, the most-fit candidate does not have a guide itself and only moves along its velocity vector in every iteration. This may yield a noticeable number of agents converge into local optima if the guide (i.e. the most fit candidate) agent cannot explore the best solution. In this paper, an attempt is made to get the advantage of the Ant Colony Optimization (ACO) methodology to assist the PSO algorithm for choosing a proper guide for each particle. This will strengthen the PSO abilities for not getting involved in local optima. As a result, we present a promising new hybrid particle swarm optimization algorithm, called ACPSO. The capability of the presented algorithm to solve a nonlinear optimization problem is demonstrated using different case studies carried out for optimal reactive power procurement. The IEEE 14-bus and 118-bus systems are adopted for reactive power market simulation. The main objective of the market is to minimize total generation costs of reactive power and transmission losses at different voltage stability margins. Based on the GAMS modeling language and the CONOPT solver, solutions are obtained for different models using a conventional non-linear optimization technique. Compared with the solutions found by the GAMS, Genetic Algorithm (GA) and the original PSO, the proposed ACPSO algorithm can provide promising results in terms of robustness and overall efficiency when it is applied to the reactive power market. [ABSTRACT FROM AUTHOR]