There are a lot of approaches for solving planning problems. Many of these approaches are based on `brute force` search methods and do not care about structures of plans previously computed in certain planning domains. By analyzing these structures we can obtain useful knowledge that can help in finding solutions for more complex planning problems. Methods described in this paper are based on analysis of action dependencies appearing in plans. This analysis provides new knowledge about the planning domain that can be passed directly to planning algorithms to improve their efficiency.