The optimal pump scheduling allows for computing the most economical energy costs and provides more efficient operations for complex water distribution systems (WDS) with multiple pumping stations. The proposed technique employs the latest advances in multi-agent Particle Swarm Optimization (MOPSO) to automatically determine the most cost-effective solutions for scheduling/operation multiple pumps in multiple pumping stations, while satisfying both loading conditions and hydraulic performance requirements. The present work considers a bi-objective pump-scheduling problem, where the objectives are: minimize the electrical energy cost ($/KW.h) and minimize the maintenance costs in terms of the total number of pump switches. In additional to the bi-objective pump-operational problem, pressure and tank levels (i.e., initial, minimum, and maximum) are considered as constraints in this paper for computing the most cost-effective solutions. The constraint-handling method, the Modified MOPSO (M-MOPSO) algorithm, and the modified EPANET Toolkit 2.0 are used to solve the constrained multi-objective problem. The results showed that the new MOPSO algorithm produced the most economical pump scheduling solutions.