According to the metal powder injection molding process, the main influence factors of injection molding billet density distribution (such as: injection velocity, injection temperature, injection pressure, etc) was analyzed and a multiple input & multiple output BP neural network model for injection molding was build up to predict the density distribution of the billet intelligently based on ANN and GA. In addition, in light of the requirements for the density distribution of the metal injection molding billet, the influence factors were controlled intelligently. Applying this model in the metal injecting process, the density distribution of billet was predicted according to the injection parameters and the injection parameters was optimized according to the required density distribution of the billet. As the result, the error was less than 5% between the prediction values and the actual values of the density distribution of billet. With the optimized injection parameters to the injection process, the density distribution of billet closed to the requirements was achieved.