By using powerful spinning SXD100/S3-CNC numerical control machine

different processing parameters（thinning ratio

feeding ratio

working angle）of connecting rod bushing manufactured and the springback were measured.Based on orthogonal experiment

taking the thinning ratio and the feed ratio

the working angle as the input layer

and the springback degree as the output layer

prediction model of powerful spinning rod bushing springback was established by BP neural network

and it was trained and predicted by testing data.The results reveal that the model exhibits a higher prediction accuracy

and it can predict accurately the springback of connecting rod bushing manufactured under different powerful spinning processing parameters.