The classic multi-objective evolutionary algorithm based on Pareto dominance criteria and crowding distance sorting method does not consider the preference of decision maker in the iterative process, which leads to the decline of convergence performance. For the problem, this paper proposes a multi-objective binary particle swarm optimization based on E-dominance to solve distribution network service restoration. By adopting E-dominance strategy of improving the origin distance, it can integrate the preference of decision makers into the evaluation process. This paper adopts roulette strategy to update the global best solution and integrated value to maintenance the external files, which can effectively guide the next generation of particle with preference of decision makers during iterative process. Finally, an example shows that the approach has better convergence performance, less quantity and better quality on solution than the classic multi-objective algorithm based on Pareto dominance criteria and crowding distance.