With the interconnection of distributed generators (DG) to conventional power system, the efficiency of distribution system management is required to improve. This paper proposes a new algorithm named as Intelligent Flow Algorithm (IFA) for reconfiguration issue, which innovate the numerical searching methods in conventional AI algorithms, to generate the optimal supplying configuration based on its topological property. The idea of IFA is based on the heuristic method to balance the distribution of loads supplying to realize low line loss. The new coding rule, named as flow generation has been developed to satisfy the generating approach, and power information besides line loss from load flow calculation, such as voltage and degree of connectivity breaking have been utilized to guide the advancing of the generating. A typical 43-bus distribution system with 20 cases is selected for case studies to demonstrate the effectiveness of the proposed algorithm. Simulation results show that IFA has high rate of global optimization within short calculation time. IFA shows stability and reliability in simulation, and the convenient adjustment of coefficient parameters also makes this method practical.