Abstract: In the study of public opinion spread, the evolution rules of the traditional cellular automaton are the majority principle which provides most emotional represents the tendency, but ignoring public opinion between individuals influence each other, it can lead to wrong prediction results, therefore, this paper combines decision theory and cellular automata and applied to the public opinion spread. First of all, introduce the basic concept and characteristics of cellular automata and decision theory. Cellular automata model based on Decision theory is set up, the algorithm of cellular state evolution is designed with relative income model as the evolution rules. Finally examples and simulation indicates that algorithm can realize public opinion spread tendency obviously, the results verify the feasibility and effectiveness of the algorithm.