Abstract: To ensure stable execution of the construction schedule and maintain low costs for extra-long railway tunnel engineering in uncertain environments, a robust project scheduling optimization model was established with dual objectives of maximizing robustness value and minimizing total cost under deadline constraints. The model incorporated resource transfer impacts on time buffers and introduced a time buffer calculation method for start-start precedence relationships considering rock grade variations. A TS-NSGA-Ⅱ tailored to model characteristics was designed, employing NSGA-Ⅱ with adaptive genetic operations as its core framework, while hybridizing a tabu search based elite strategy for constructing neighborhood solutions to enhance both global and local search capabilities. Case simulations demonstrate that the proposed model and algorithm can effectively obtain stable scheduling schemes with low construction costs, and the robustness value improves by 11.33% compared to robust project scheduling optimization model that does not consider resource transfer. The comparison of case tests show that the TS-NSGA-Ⅱ is superior in terms of solution quality and solution efficiency, and can meet the scheduling optimization requirements for extra-long railway tunnel engineering in uncertain environments.