The isothermal CVI (chemical vapor infiltration) process of carbon/carbon composites is controlled by many factors and its efficiency is very low. Manufacturing of C/C composites with isothermal CVI processes is costly and thus limits commercial applications of C/C composites. Based on the fuzzy neural network (FNN) technique and genetic algorithm, a predicting system for isothermal CVI process was proposed and established. The simulation results of FEM called virtual samples were selected as the network's training samples. Based on the FNN system, the influences of main infiltration parameters, such as infiltration temperature, precursor gas flow ratio and rate, had been studied; and they are good instructions for the design and optimization of CVI process. Using this FNN system, we expect that we can reduce the time of development and densification, thus reducing manufacturing cost.