This study is analyzed variations of ultimate strength, friction coefficient and wear mass loss that depend on the injection molding techniques during the blending of short glass fiber (SGF) and polytetrafluoroethylene (PTFE) reinforced polycarbonate (PC) composites. A hybrid method including response surface methodology (RSM) and back-propagation neural network (BPNN) integrating simulated annealing algorithm (SAA) are proposed to determine an optimal parameter setting of the injection molding process. The specimens are prepared under different injection molding processing conditions based on a Taguchi orthogonal array table. The results of eighteen experimental runs were utilized to train the BPNN predicting ultimate strength, friction coefficient and wear mass loss. Simultaneously, the RSM and SAA approaches were individually applied to search for an optimal setting. In addition, the analysis of variance (ANOVA) was implemented to identify significant factors for the injection molding process parameters and the result of BPNN integrating SAA was also compared with RSM approach. The results of optimal parameters of injection molding process for the ultimate strength of x-direction and y-direction based on BPNN/SAA approach were increased 3.12%, and 6.18%, respectively.