In this paper, a stochastic homogenization method that couples the state-of-the-art computational multi-scale homogenization method with the stochastic finite element method, is proposed to predict the statistics of the effective elastic properties of textile composite materials. Uncertainties associated with the elastic properties of the constituents are considered. Accurately modeling the fabric reinforcement plays an important role in the prediction of the effective elastic properties of textile composites due to their complex structure. The p-version finite element method is adopted to refine the analysis. Performance of the proposed method is assessed by comparing the mean values and coefficients of variation for components of the effective elastic tensor obtained from the present method against corresponding results calculated by using Monte Carlo simulation method for a plain-weave textile composite. Results show that the proposed method has sufficient accuracy to capture the variability in effective elastic properties of the composite induced by the variation of the material properties of the constituents.