The carbon emission of the power industry is the main component of China"s carbon emissions. In order to achieve the double carbon goal, the new low-carbon power system has become the inevitable trend of the future development of the power system. The power system with a high proportion of new energy has weak immunity and low overload capacity. The starting of large induction motors may lead to equipment disconnection, which will adversely affect the stability of the system. Therefore, the efficient and accurate calculation of motor starting process is particularly important for the new power system transient time domain simulation. In order to obtain an efficient stable algorithm for induction motors during starting process that can interface with common transient simulation programs in power systems, and thus improve the efficiency of transient simulation, this paper compares the accuracy, stability and efficiency of four algorithms: explicit Euler, implicit Euler, implicit trapezoid, and Runge Kutta mehod, and applies the above algorithms to the simulation analysis of the starting process of high-voltage high-power induction motors. From the point of view of numerical iteration, the causes of oscillation in the steady state of different algorithms using large step size simulation are analyzed. In order to solve the numerical oscillation problem, variable parameters are introduced into the traditional implicit trapezoid method to obtain an improved trapezoid method, and the optimal parameter interval for reliable convergence of the algorithm is determined. Finally, the effectiveness and feasibility of the improved trapezoidal method are verified by numerical simulation, which provides guidance for the selection of efficient simulation algorithms in new power system.