In the current research of vacuum arc brazing technology, there are problems that the key factors of arc ignition process are not clearly defined and the value of the factors is adjusted mostly based on experience. In this paper, the influencing factors of arc ignition process of vacuum arc brazing equipment are studied. To quickly build arc and get stable arc current, argon flow, breakdown voltage, breakdown distance and tantalum tube diameter cited various influential factors in the process of arc are analyzed one by one from the establishment of the basic mechanism. The optimal parameter prediction model is established based on adaptive neural network based fuzzy interference system(ANFIS) model. The argon flow, vacuum degree, the breakdown distance and the diameter of the tantalum tube as the prediction model input are used to predict the breakdown voltage optimal value. Through the field experiment, the theoretical research and the prediction performance of the parameter prediction model studied in this paper are verified and improved. The results show that the arc current in arc ignition process of vacuum arc brazing equipment is stable, which uses the parameters obtained from prediction model, and results meet process requirements.