Rules are widely used in artificial intelligence for modeling intelligent behavior and building expert systems. Most of the rule-based system programs are extremely computation- intensive and run quite slow. This problem may get worse when a rule-base does not fit into primary memory completely. On the surface, rules appear not fit for indexing unlike the databases where indexing looks obvious. In this paper, we present two sources of indexing— attribute level and value level indexing— which improve the run time response of the rule-based systems. A forward chaining inferencing using the indexing method proposed is also presented. Simulation results show that indexing in rulebases improves run time efficiency of large scale knowledgebased systems.