Alumina ceramic coating was gained by micro-arc oxidation of hot-dip aluminum film on stainless steel. The main properties indexes of ceramic coating include the thickness, the hardness and the surface roughness. The main process parameters determine above properties include the electric current density, the duty ratio of the pulse, the frequency of the pulse and the action time. The artificial neural network model was established to forecast the properties and optimize the process parameters for the micro-arc oxidation of aluminum film. Experimental results proved that the optimization model has high veracity and can be applied to the choice of process parameters.