Water network rehabilitation is a complex problem, and many facets should be concerned in the solving process. It is a discrete variables, non-linear, multi-objective optimal problem. An optimization approach is discussed in this paper by transforming the hydraulic constraints into objective functions of optimization model of water supply network rehabilitation problem. The non-dominated sorting Genetic Algorithm-II (NSGA-II) was adopted to solve the altered multi-objective optimal problem. The introduction of NSGA-II for water supply network optimal rehabilitation problem results in solving the conflict between one fitness value of standard genetic algorithm (SGA) and multi-objectives of rehabilitation problem. Moreover, it benefits to control the uncertainties brought by using weighting coefficients or punish functions in conventional methods. In order to accelerate the convergence speed of population, this paper introduces the artificial inducement mutation (AIM). It not only improves the convergence speed, but also improves the rationality and feasibility of solutions.