The reasonable path planning of robot or manipulator in the workspace is the key to improve the search efficiency and reduce the search distance.However, it is more important to ensure the safety of its work process. The existing path planning algorithms all have problems such as complex search process and insufficient security considerations. In this paper, an improved A-star optimal path planning algorithm is proposed, which produces safety-insured optimal path that are further fine-tuned with smoothing turning corners. First, a comparison study is carried out for a number of neighborhood searching algorithms in the family of the A-star approach. It is found that the simplest four neighborhood A-star searching algorithm is often most practical. Next, we introduce a matrix in the A-star algorithm, that ensures safety distances for operations along the optimal path. In addition, the heuristic functions used in A-star algorithms are  modified to ensure the optimal safety path under different environments. Finally, a path trajectory optimization method based on different environmental conditions is proposed to ensures safety distances. The results show that the improved A* algorithm proposed has few path turning points, short search time, simple post-processing, and can ensure the safety of operating conditions along the path.