Abstract: Dynamic measurement error always exists in the measurement system, so this paper proposes a method for virtual instrument dynamic compensation of measurement system based on RBF neural network to eliminate the dynamic measurement error and reappear the primitive input signal which is measured accurately.This method uses the strong nonlinearity approximation ability of neural network to obtain the compensating system model parameters according to the measurement system′s input and response data, and constructs the dynamic compensating system of measurement system through LabVIEW, which does not rely on the mathematical model of measurement system.The test result indicates that it is effective to carry on the dynamic compensation of measurement system by unifying the RBF neural network and the virtual instrument.