Abstract: The application of active control and management measures to active distribution network (ADN) can effectively solve the technical problem of distributed energy connected to the grid with high permeability. With respect to the comprehensive optimal allocation problem of ADN connected with multi-type distributed generation (DG), energy storage and compensation capacitor, a multi-objective optimization model is built by considering the integrated cost, average power reliability and emissions of greenhouse gases, and a kind of improved multi-objective particle swamp optimization algorithm is proposed to solve this problem. This algorithm can improve the global optimization ability of the algorithm and avoid premature by introducing genetic operators, and can improve the capacity of optimizing capacity allocation of multi-objective particle swarm optimization algorithm through the external file updating and selecting strategy of group global optimal solution in dealing with capacity allocation problem for DG, energy storage and compensating capacitor. For the obtained non-inferior solution set, a fuzzy decision technique based on the information entropy is applied to obtain the final scheme, which is immune to the preference of decision-maker. The IEEE-33 buses distribution system is chosen as simulation object to validate good converge property and global searching ability of the algorithm.