Artificial Neural Network ( ANN) is a system of analysis which it’s process inspired by activity of neuron network at human being. This network is consisted of a group of neurons or units which interact each other. At neuron of human, natural process arrange how input signal at dendrite processed and then translated into axon activity. While at neuron of made in, learning process arrange input that used for its output mapping. This research will be applicated the job process of ANN by grouping growth of a baby in womb into 3 different phase, that is pre-embryonic phase, embryonic phase and foetus phase (foetus). Where the third of phase can apart well and also can determine distinguishing function between goups and classify new object into existing group or class. ANN is one of alternative solving in discriminant analysis with unlinear function. The advantage of this method is assumption needn't for data that have to distributed normal multivariate and having high correctness. From result of data analysis, seen that by using ANN method of third phase of fetus forming that is pre-embryogenic, embryogenic and foetus phase, got a most stable model is 14 hidden node (one hidden layer) and will give subdividing mistake most minimum (mean square error) equal to 0,0396941 and conditional opportunity probability (proportion correct) equal to 100%.