The paper establishes a mathematical model of distribution network reconfiguration with minimum network loss as objective function. An improved genetic algorithm is used to seek the best reconfiguration scheme of distribution network with distributed generations and energy storing devices. The algorithm adopts a decimal encoding, which regards the number of loops as the length of the chromosome and the on/off states of the switches in loop network as the chromosome genes. The distributed generations and energy storing devices are equivalent to "negative" load in calculating power flow. Based on the unique hierarchical structure characteristics of the distribution network, the paper applies the layer-by-layer back/forward sweep method to solve load flow. Simulation result of IEEE33 node example indicates that the reconfiguring the distribution network with distributed generations and energy storing devices can decrease network loss and voltage drop to a large extent. This work is supported by Science and Technology Project of State Grid Corporation of China (No. 5216A016000K and No. 5216A2160004).