To the problem of corresponding points matching inaccurately using linear feature generalization method with multi-scale map data, which causing poor spatial data quality of intermediate scale linear feature generated by interpolating, a corresponding point matching method was proposed based on the Optimum Correspondence(OptCor) algorithm. In the method, the corresponding points were identified by establishing the most excellent corresponding relationship among the linear features in different scale map data, which avoiding the situation of being difficult to match with many-to-one, one-to-many, many-to-many, or one-to-one, but not being the best match using buffer method. And intermediate scale linear features were generated through multi-scale linear feature generalization interpolation using the method. Experiments were conducted with three types of line element data.The experimental results show that the multi-scale linear feature generalization method based on OptCor algorithm is more accurate and efficient in matching corresponding points than buffer method,and generalized results of the method are closer to real data.