Technological forecasting for microprocessor market can provide timely insight into the prospects for significant technological changes in computer hardware as well as software. In this paper, we use bibliometrics to forecast R&D trend on microprocessor technology. Cumulative numbers of US Patents on several generations of microprocessor technology (pipeline, superpipeline, supersclar and VLIW) approved since 1980 are applied to fit diffusion models. Our study results provide both the maximum market potential and the maturity time for each generation of microprocessor technology. Such information is expected to make contribution on making better decisions with regard to strategic corporate planning, R&D management, product development and investment in new technology of microprocessor.