Short-term and (Mid\Long)-term load prediction are needed for the efficient management of distribution system. An accurate load forecasting could help distribution operators to achieve a great control and flexibility on energy. In addition, a reliable prediction would have significant impact on load power flow problems, operation management, planning, automation, and search market. This paper proposes a new method based on copula for Short-term, and (Mid/Long)-term load forecasting. Proposed method was used less historical data to predict the load consumption then the other methods, so it is useful in distribution system due to scarce installed meters. Copula is a recently developed statistical theory for multivariate probability analysis that has been proposed by Sklar in 1958. Copulas are used in popularity for financial and reliability application. Numerical tests results using database of Iran system’s operation (ISO) shows that this method provides a much better prediction performance in comparison with other methods employing the same data.