Nowadays, natural computation has been widely used to solve for complex multiobjective optimization problems in diverse areas. Enlightened by the membrane computing and the mechanisms of fireworks explosion in the sky. A hybrid components of multi-objective optimization algorithm, called membrane-fireworks algorithm (MFWA)was proposed. In order to balance the capability of global and local searching in algorithm, the fireworks explosion optimization approach is used in elementary membrane to improve the local searching ability and the elite opposition-based learning mechanism is adopted for global optimization. Moreover, aiming at the improvement of searching efficiency, the crowding distance and non-dominated sorting are used to update multiple sets in skin membrane. The results of simulation experiment show that the MFWA is feasible, effective and capable of achieve the better result sets closer to Pareto front on typical benchmark set including six ZDT and DTLZ series test functions.