The management of resources and task scheduling in a distributed computing is complex and challenging due to various factors such as the resource’s geographic distribution, heterogeneity, dynamic, distributed ownership with different policies and priorities, varying loads, reliability, and availability conditions. Performance measurement, analysis and prediction of distributed systems have become increasingly important. In particular, accurate load prediction of each computer is crucial in achieving required performance for practical applications. This paper presents a resource schedule algorithm based on its reliability, which is the centre of any scheduling and resource allocation algorithm. The algorithm can effectively resolve the problem of load balancing in a distributed system.