Abstract: Robotic assembly is a complicated dynamic process. High speed assembly may incur serious damage to robotic systems and workpieces. In order to find out an effective method to resolve the problem, this paper presents a progressive learning strategy of RBFN to learn the dynamics of the assembly process. We then implement an intensified learning algorithm using gradient search method to adjust impedance components. All of the strategies are based on the dynamic equations that are derived on the basis of the concept of impedance control. Computational simulation results are given to demonstrate the effectiveness and advantage of the approach.