Introduction Currently, there are difficulties in dealing with higher construction requirements and standards in subway construction management. Therefore, a multi-objective optimization model was constructed based on building information management technology, and an improved non-dominated sorting genetic algorithm III was introduced to optimize the model solution. And experimental verification was conducted. These experiments confirmed that the average HV of the improved algorithm was 0.67, which was higher than the original algorithm's 0.65, indicating that it had higher convergence and reliability. The solution results of the non-dominated sorting genetic algorithm II showed that the optimized cost was 185.1899 million yuan. The cost of optimizing the original non-dominated sorting genetic algorithm III was 184.6469 million yuan. The total cost of optimizing the research algorithm was 184.1165 million yuan. In addition, the research algorithm had the shortest construction period, ideal cost, and significantly higher quality and safety levels than the comparison algorithms. And its time consumption was only 20 seconds, significantly lower than the comparison algorithms. And its cost was between 183 million to 187.5 million yuan, with higher stability and relatively concentrated distribution of solutions. Overall, the subway construction optimization model based on building information management and non-dominated sorting genetic algorithm III has high effectiveness and can be effectively applied in practical construction management. Material and Methods Results Conclusions