Abstract: To address the problems commonly encountered by tower crane groups in large-scale construction projects, including limited site space, operational interference among work zones, and high potential collision risks, as well as the difficulty in simultaneously balancing operational efficiency, cost, and safety in their layout, an improved NSGA-Ⅱ-based multi-objective optimization method for tower crane group layout in large construction sites is proposed. Firstly, a multi-objective optimization model for tower crane group layout is established, with the objectives of maximizing the operational efficiency equivalent, comprehensively characterized by average operational efficiency and operational coverage rate, while minimizing the cost equivalent and safety equivalent, with the latter quantified by the degree of overlapping coverage. Secondly, to improve the quality of the initial population of the NSGA-Ⅱ algorithm under complex constraints, a global candidate-point pool based on Latin Hypercube Sampling (LHS) is constructed, and a hybrid initialization mechanism combining random subsampling and farthest-point sampling is designed. Finally, to balance global exploration and local exploitation during the solution process, a dynamic adaptive crossover operator based on parental quality and an adaptive mutation strength strategy are designed, while a periodic random immigration mechanism is introduced to maintain population diversity. Experimental results demonstrate that, compared with the standard NSGA-Ⅱ, the proposed method achieves faster convergence and better solution set quality. The optimized layout scheme yields effective improvements in all three objectives, namely the operational efficiency equivalent, cost equivalent, and safety equivalent. Meanwhile, the coverage rate of the tower crane group increases from 97.30% to 98.32%, high-order overlapping areas are effectively avoided, and the second-order overlapping area is reduced by 70.28%. It achieves a better comprehensive balance of efficiency, cost and safety, providing an effective intelligent optimization approach for tower crane group layout in large construction sites.