Research and development (R&D) project termination decision is an important and challenging task for organizations with R&D project management .Current research on R&D project management mainly focuses on project selection decisions. Very little research has been done on the termination decision of R&D projects .In this paper a support vector machines classifer for assisting managers in deciding whether to abandon an ongoing R&D project at various stages of R&D is presented. It has also shown by the modeling and pattern recognizing results in terms of termination decisions of fifty R&D projects that the method possesses reinforcement learning properties and universalized capabilities. With respect to modeling and termination decision of R&D project, which has the fact that the evaluation criteria are hardly ever determined by conventional approaches such as statistical analysis, the method is available.