Twodimensional cellular automata model is often applied to simulate the forest fire spreading. However, the model has many disadvantages which include numerous iteration times, unfinished evolutionary process, uncertain condition of iteration, etc. Aiming at tackling these problems, a 3D cellular space fire spread model based on genetic algorithm with multiple objectives is proposed. First, the main factors that influence the spreading of forest fire are applied to the twodimensional cellular automaton model. Secondly, to decrease the iterations of cellular automaton model of spreading forest fire and identify the termination condition of iteration, the initial mode of the twodimensional model is improved which utilizes the threedimensional spherical cell space. The genetic algorithm with multiple objectives is applied to cellular automaton algorithm for improving the prediction accuracy of forest fire spreading model. By comparing the proposed method with the traditional twodimensional cellular automaton model, Wangzhengfei model and Rothermel model, it can be seen that the proposed method in this paper can greatly reduce the number of iterations and the running time. The operation efficiency of the cellular automata is also improved remarkably. In addition, the termination condition of cellular automata is explicit. The actual forest fire spreading process has high similarity with the simulation results of the proposed forest fire spread model.