This paper proposed a multi-objective evolutionary algorithm (called by GDE-EDA hereinafter). The proposed algorithm combined a generalized differential evolution (DE) with an estimation of distribution algorithm (EDA). This combination can simultaneously use global information of population extracted by EDA and differential information by DE. Thus, GDE-EDA can obtain a better distribution of the solutions by EDA while keeping the fast convergence exhibited by DE. The experimental results of the proposed GDE-EDA algorithm were reported on a suit of widely used test functions, and compared with GDE and NSGA-II in the literature.