Proteins are amino acid chains that acquire their biological and biochemical properties by folding into unique 3-dimensional structures. The biological function of a protein is dependent on the protein folding into the correct, or "native", state. At present, there are so many ideas to predict the structure of the protein folding. This paper first present the concept of protein folding and how is significant to study protein fold prediction. In this paper we join the simulated annealing factor into Parallel Genetic Algorithm and use this hybrid Parallel GA to predict the structure of protein fold. The revised algorithm is more efficient than traditional Genetic Algorithm and simulated annealing algorithm.