To meet the demand of working under high speed, heavy duty conditions, and low noise emission, high precision gears are required in geared power transmission systems of high-end equipment. In order to improve the gear precision as well as to lower the gear manufacturing cost, in this study, it is aimed to establish the mapping rules between the gear hobbing processing technique and gear geometric errors. Based on the control variable method, the key gear processing parameters, such as cutting speed and feed rate, are selected as the input parameters for the proposed model. The output parameters are the gear geometric errors obtained by using the precision gear measurement machine, which include total error of the tooth profile, total helical error of the tooth surface, single pitch error of the tooth surface, and accumulated pitch error of the tooth surface. These data are recorded and analyzed by applying the back propagation neural network algorithm that predicts the aforementioned errors of interest if certain input parameters are provided. Furthermore, the predictions made by using the proposed model are validated using experimental results. The accuracy of the proposed model is evaluated using the root meansquare error equation that quantifies the predicted and true values of gear geometric errors. The analysis method reveals the mapping rules between the gear hobbing processing technique and gear geometric errors, and thus can provide guidance for optimization of gear’s processing parameters for precision gear manufacturing.