Since cables laying in the tunnel have high load capacity and convenient maintenance,they have been widely used in the urban power grid. In order to make full use of transmission capacity,a phase sequence optimization method of tunnel multi-circuit cables based on an improved genetic algorithm and multi-physical field simulation is proposed. Firstly,considering the electromagnetic environment,heat dissipation environment and the cross-bonding grounding mode,an electromagnetic-heat-flow-circuit multi-physical field coupling model of tunnel cables is established. Secondly,the memory population is set up in the genetic algorithm to lessen the invocations of the finite element model and improve the calculation efficiency. Finally,an improved genetic algorithm is applied to the phase sequence optimization of cables in which minimizing the maximum temperature of cable cores is considered as the objective function and the cable temperature of each loop not exceeding 90 ℃ as its constraint condition. After optimization,the total sheath circulating loss of the cables decreases by 11.14%,and the maximum temperature of the cores decreases by 2.15%. The results show that the optimization of phase sequence is conducive to improving the heat dissipation of the tunnel cables,and the improved genetic algorithm greatly decreases the optimization times.