We present a novel approach to fuse the Radius Particle Swarm Optimization and Simulated Annealing (RPSO-SA) to solve the Knapsack Problems (KPs). The features RPSO-SA create an innovative approach, which can generate high-quality solutions in shorter times and more stable convergence characteristics. The RPSO takes advantage of group-swarm to keep the balance between the global exploration and the local exploitation. The SA gently improves the candidate solution by searching for optimal solutions within a local neighbourhood. The RPSO-SA combines the strong global search ability of RPSO and the strong local search ability of SA to reach faster optimal solution. In addition, there are two ways of accepting a new solution. The method has been tested against the knapsack problems. The results indicate that the combined approach outperforms individual implementations of radius particle swarm optimization and simulated annealing.