In this paper, a multi-product capacitated newsvendor model with stochastic demand is studied. Firstly, a multivariate Markovian demand model to forecast the demands of multiple products is constructed, and the relationships among the demands are measured by the model. Secondly, a capacitated newsboy model with multivariate Markovian demand is proposed, and then the solution of optimal ordering policy for multiple products is derived. At last, the properties of the optimal policy through the probability value and truncated probability of the expected demand state are discussed. With capacity constrains, the theoretical analysis shows that the optimal ordering quantity is a monotonic decreasing function with respect to the left truncated probability, and a monotonic increasing function with respect to the right truncated probability, but the probability value of the expected demand state has special dual effects on the optimal ordering quantity.