Abstract: A modified attractive and repulsive particle swarm optimization (MARPSO) algorithm is proposed, which is based on the attractive and repulsive particle swarm optimization (ARPSO) algorithm, by employing new diversity-measure and repulsive operator. Combining both the local convergence ability of hill climbing(HC) algorithm and the characteristic avoid precocious of MARPSO, the way of automatic generation of the software test data based on hill climbing algorithm combined with MARPSO(HC-MARPSO) is proposed. Finally, the results of experiment show that this new algorithm can generate test data more effective than other algorithms, such as genetic algorithm and particle swarm optimization algorithm.