The dynamic error of CMM measurement process restricts the improvement of measurement efficiency in industrial field. For this purpose, a dynamic error compensation method for high-speed measurement process of CMM is proposed to improve measurement accuracy and efficiency. In order to make the research typical and representative, the mobile bridge CMM widely used in the market is taken as the research object. By establishing an error separation platform to analyze the dynamic characteristics of the high-speed measurement process of the measuring machine, four parameters that can characterize the dynamic errors of the measurement process are determined, namely maximum positioning error (MPE), residual positioning error (RPE), maximum approximation error ( MAE), residual approximation error (RAE). Orthogonal experiment method was used to analyze the influence degree of common influence factors (positioning speed, positioning distance, approaching speed, approaching distance) on dynamic parameters. Training samples and test samples were obtained by measuring standard spheres with coordinate measuring machine. The training samples and test samples are used to model and compensate the dynamic errors of the measuring machine. The results show that the dynamic process errors are reduced by 88. 8%, 80. 2%, 90. 8% and 71. 3% respectively after compensated by the fuzzy neural network model. It proves that the fuzzy neural network model can effectively improve the dynamic measurement accuracy of the measuring machine.