Abstract: Dynamic multi-objective optimization problems (DMOP) are some problems whose objective functions, constraints, or parameters change dynamically. DMOP are important and challenging tasks in the real-world optimization domain. Generally, it is difficult to track the Pareto front of DMOP by the traditional evolutionary algorithm. Aimed at the characteristic of dynamic multi-objective problems, a novel co-evolutionary algorithm for DMOP (dynamic multi-objective optimization algorithm based on ecological strategy, ESDMO) is proposed based on ecological strategies and a new self-detecting environmental change operator. The ecological strategies are very important for individuals to fit to the changing environment and get higher competition ability. The proposed method adopts an evolutionary computing model that combines co-evolution mechanism and reinforcement learning strategy, which is inspired from ecological strategy between predator populations and prey populations. A self-detecting environmental change operator is defined and used to measure the changing environment in the algorithm. Hereby different populations take different ecological strategies to cope with environmental change. Several typical dynamic multi-objective problems are tested. The experimental results show that the proposed algorithm can get better diversity, uniformity and convergence performance. It demonstrates that the proposed algorithm is effective for solving DMOP.