This paper introduces an improvement approach of the classical backtracking algorithm tailored for solving complex combinatorial problems. The proposed approach restructures the search space by grouping solution sets into meta-sets and applying continuity criteria early in the search process, leading to a significant reduction of the number of iterations required to reach valid solutions. The effectiveness of the approach is demonstrated through case studies involving Sudoku and Kakuro puzzles, where it outperforms traditional backtracking in terms of computational efficiency and solution accuracy. The results show that the proposed approach offers a scalable and practical enhancement for combinatorial optimization tasks.