In this paper, an improved Simulated Annealing algorithm for Protein Folding Problem (PFP) is presented. This algorithm called Cluster Perturbation Simulated Annealing (CPSA) is based on a brand new scheme to generate new solutions using a cluster perturbation. The algorithm is divided into two phases: Cluster Perturbation Phase and the Reheat Phase. The first phase obtains a good solution in a small amount of time, and it is applied at very high temperatures. The second phase starts with a threshold temperature and reheats the system for a better exploration. CPSA reduces the execution time of the Simulated Annealing Algorithm without sacrificing quality to find a native structure in PFP in Ab-Initio approaches.