Abstract: In the present study surface roughness prediction model for end milling of Al-SiCp metal matrix composites using artificial neural network is developed. The SiC percentage in the metal matrix composites is considered as one of the input parameters in addition to three prominent variables i.e., spindle speed, feed rate and depth of cut for predicting surface roughness being the output variable. A multi layer perception network having 4-14-1 architecture is found to be optimum network. The network is trained with 25 data sets and the trained network predicts the surface roughness for the interpolated values of input parameters. The result shows that the prediction performance of the neural network model is highly encouraging with an average percentage error being 0.31%. While response surface model predicts an average percentage error of 0.53%. The surface roughness is mainly affected by feed rate followed by other parameters spindle speed, SiC percentage and depth of cut.