Digital design offers sustainable fashion design the possibility of continuous development. This paper proposes an intelligent clothing generation framework that integrates the improved iterative closest point (ICP) algorithm with a particle-spring physical model to achieve precise virtual generation of fashion clothing. The ICP algorithm is used to preprocess human point cloud data, removing incorrect corresponding point pairs, and combining four types of constraints to improve the accuracy of coordinate transformation. A particle-spring model is constructed using fabric dynamics simulation to simulate the effects of particle forces, enhancing the realism of clothing design. Experimental results show that the average error in generating clothing components is below 1.5%. The time required for model adjustment operations for three different types of clothing is only 9, 8, and 10 seconds, respectively, with post-adjustment errors below 0.70%. The trajectory error of the generated clothing model is less than 0.30, with maximum average curvature and average position errors of 19.6% and 2.649 units, respectively.
