Abstract: This paper introduced Artificial Neural Network(ANN) into vibration drilling. The model and algorithm of ANN that were applied to varying-parameters vibration drilling simulation and parameter optimization were mainly studied. In this paper, the orthogonal test method was used to organize training samples. Because of its lack, K-dimension crossing test method was also used. There were 27 groups of samples, which were divided into 9 training groups. In training processing. The different samples were input in intersection way, and one of the samples was input as different as possible from another. Network was trained for 9 times, and every time one subgroup was out of training, the selected one was used to test generalization errors. The experiments showed that the ANN systems are of high precision. It also provided a new method for vibration drilling studying and analysis.