Abstract: The effect of void distribution, size and shape on transverse tensile strength of unidirectional carbon fiber-reinforced epoxy resin composites with same void content was studied. SEM images of composites were processed using Matlab and the void radius distribution of composites was obtained. The void random algorithms were programmed by C++, including void random distribution algorithm, different void radius algorithm and different void shape algorithm. The representative volume element(RVE) of the composites microstructure was generated by Python. The effect of void distribution, size and shape on transverse tensile strength of unidirectional carbon fiber/epoxy resin composites was simulated by the finite element analysis. It is found that carbon fiber/epoxy resin with the same void content, the effect of void shape on transverse modulus is relatively large, and the void size and void shape affect the transverse tension strength largely.