Abstract: For optimizing the cotton blending problem, an improved differential evolution algorithm is proposed. Firstly, a constrained optimization model for cotton blending problem is built. Then the differential evolution algorithm is improved to solve this model, which uses the penalty function to handle multi-constraints, and dynamically adjusts control parameters in the evolution process. Finally, the experiment is carried out to verify the proposed method using the actual raw cotton data from a textile enterprise. The results show that this method is superior to classifying and queuing method and genetic algorithm, which could effectively reduce cotton blending costs for an enterprise, with a high practical application value.