A novel adaptive control strategy is designed for MOEA/D. The parameter settings are adaptively adjusted at different evolutionary stages. The advantages of four classical differential evolution strategies are combined. Our algorithm gives better performance in solving complicated UF test problems.