Resource request task is limited by scheduling length, time span, load balancing, quality of service, economic principles, safety performance and other factors in the Enterprise Gloud Storage system, it is difficult to characterize it with specific features. This paper analyzed the storage model, designed a compromised task scheduling algorithm based on multi-target balance. Algorithm�s universality is verified by large-scale measurements which based on the randomly generated maps. The experimental results indicate show that in view of the computation intensity and the memory intensity load, this algorithm can perform well on aggregative indicators such as job scheduling length, scheduling cost and average energy consumption.