History matching is a long-standing problem in reservoir simulation. Although automatic history matching is a difficult process due to the large number of variables and high degrees of uncertainty, it is possible to develop methodologies to improve the matching process. This paper presents MASTER Web, a software tool for configuring a web-based parallel computing system on a cluster of PCs to support a new approach for history matching. Simple Java-based software has been implemented to connect/communicate heterogeneous (different hardware and operating system) PCs for carrying out the matching process. MASTER Web, a loosely coupled parallel system (with no shared memory) configured in a master-slaves fashion, is sufficient for history matching since there is no need for inter-slave-processor communication. Utilizing parallel computing across the network (the Web), this novel technique not only aims to improve history matching, but does it economically on clusters of ordinary PCs, making it affordable for smaller oil companies. There are a great number of applications of this technique in reservoir simulation. Every application that requires a large number of repeated simulation runs can be improved with MASTER Web. The results we obtained are highly encouraging and indicate the great potential of our approach. Our experiments, carried out with a small cluster of PCs, have achieved a significant practical speedup in terms of wall clock time.