The purpose of this study is to build a model that can estimate covariances of returns easily and directly. As we integrate the covariance model and the average correlation model, the average covariance model that can estimate covariances of returns accurately and cost-effectively is built. By the way, the model assumes that the covariances in a homogeneous group is equal to averaging the elements of the group in the full variance - covariance matrix. As a result of empirical study, we could not argue the superiority of our model to the covariance model and the single index model. Reducing inputs required for the Portfolio analysis, our model will be frequently used by the investors.