Focusing on the balance between the best-fit method and the evaluation of packing layout, a new effective best-fit heuristic algorithm for 2D packing problem was proposed based on forbearing stratified strategy for multi-objective optimization. Firstly, the packing space and fit values were defined, and the wide and high fit values of the current part and the packing space were calculated, then the unified multi-objective optimization function model and packing priority rule were built based on the objective function values. Especially for the general placeable-fit situation, the best layout was finally achieved by setting and adjusting the tolerance values. The computational results on benchmark problems of 7 kinds of data show that the average Gap is reduced by 2% compared with Lowest-Level Left Align Best Fit (LLABF) and Lowest Skyline Best Fit (LSBF); the packing heights reach 24 and 339 respectively for two sets of random data of C1P1+C3P1 and C2-C7 (The number of rectangles is 33 and 66). This algorithm can also be used for irregular part packing optimization.