Existing PSO based task scheduling algorithm in cloud storage system is fail to aware users' QoS preference,namely they use the same fitness function and result in low user satisfaction rate. This paper redefined the fitness function of the PSO which was calculated by multi-QoS factors. And analytic hierarchy process was integrated to define different weights for different QoS to adapt the QoS requirement of various tasks. Simulation results show that the algorithm can adapt to various QoS requirement. Alth...