In this paper, we present a distributed computing scheme for examination feedback analysis based on grid. The proposed scheme is motivated by the observation that as the number of examinations increases, the feedback analysis would become extremely slow if a single server is used. Grid computing can be used to solve the problem by evenly distributing the workload of the system to all the available nodes, thus greatly enhancing the performance of feedback analysis in a learning system. We use Hyper Grid Learning System (HGLS)-a learning system with examination service-we developed as an exampIe to describe the proposed scheme. We note that although discussed in the context of HGLS, the proposed scheme can be easily adapted and applied to all the e-Iearning systems based on grid computing. To evaluate the performance of the proposed scheme, we compare the performance of the proposed scheme with single node servers. Our experimental result shows that the proposed scheme can greatly enhance the performance of the feedback analysis in a learning system base on grid.