Rock is a natural heterogeneous material and presents complicated behaviors in the fracturing process. It is prevail to study the basic failure mechanism of rocks via numerical simulation. Based on the elasto-plastic cellular automaton (EPCA) model, this paper simulates single pre-fractured rock fracturing process with consideration of rock heterogeneity on the meso-scale. In this model, the Weibull’s distribution, which characterizes heterogeneity with the homogeneous index m and the random seed parameter s, is adopted to describe the distribution of mechanical parameters of rock specimens such as cohesive strength, Young’s modulus, etc. Pre-existing crack rock specimens with different homogeneous index or the different random seed are simulated by EPCA under uniaxial compression. Numerical results show that heterogeneity has great influence on pre-fractured rock failure process, final failure modes, and the uniaxial compressive strength.