On the basis of the experimental analysis between peg’s assembly force and position

which is in the process of robot peg in hole assembly process

a fuzzy CMAC artificial neural network system which can be used in robot active assembly task is developed.Fuzzy CMAC artificial neural network which based on takagi type fuzzy reasoning method is adopted

and it’s learning algorithm is analyzed.With the wrist force sensor

the robot active assembly experiments are processed on the assembly robot which is five degree of freedom.The peg-hole assembly task with the diameter of 12 mm and the clearance of 15μm is realized.The experimental result shows that it is a new and effective method in robot active assembly process.