The invention relates to a method for automatically searching assistant decisions for transient and dynamic stability based on parallel computation. Based on power flow and stability data of a power grid and aimed at the problems of transient stability and dynamic stability, the method adopts large-scale parallel computation processing technology as well as human experience and a specific computation method to realize automatic identification of adjusting elements and automatic search of feasible adjustment methods, so as to achieve off-line transient stability analysis of an auxiliary electric power system and actual operation instruction.