Grid computing is currently an important research area and task scheduling is a basal part of it. The performance of task scheduling directly affects grid QoS. A task scheduling algorithm based on multidimensional performance clustering of grid service resources, MPCGSR (task scheduling algorithm based on multidimensional performance clustering of grid service resources), is proposed for shortening the completion time of task scheduling and improving task scheduling performance. In the algorithm, combined with the theory of small world, the multidimensional performance clustering of service resources is executed in advance based on the hypergraph model of grid service resources constructed according to characteristics of grid resources such as its huge numbers, heterogeneity and multiplicity. Tasks are matched to clustering resources and scheduled. Simulation results show that it is an effective grid task scheduling algorithm that is superior to other kindred algorithms.