Aiming at the problems of high cost in manual scheduling and low vehicle utilization rate for trackless rubber-tyred vehicles in underground coal mines, an intelligent scheduling model was established to coordinately optimize transportation efficiency and environmental performance. Firstly, the MT-CVRPTW model integrating hard time windows, cargo taboos and mileage constraints was established. Secondly, a bi-objective optimization model with the objectives of minimizing the total cost and carbon emissions was constructed, and a Parallel Genetic Algorithm-Adaptive Large Neighborhood Search (PGA-ALNS) hybrid optimization framework based on PGA and ALNS was proposed. The results of the simulation experiments show that, through the two-layer collaborative mechanism of the global search of PGA and the dynamic neighborhood optimization of ALNS, combined with the time-sensitive weight strategy and elite retention mechanism, the algorithm keeps the carbon emission cost at the minimum. Compared with the Genetic Algorithm and Simulated Annealing Algorithm, its total transportation cost is reduced by 11.1% and 7.6% respectively, the total distance is reduced by 10.7% and 4.7% respectively, and the solution time is shortened by 40.2% and 36.1% respectively. For the scheduling scheme optimized by the PGA-ALNS algorithm, its average vehicle load rate reaches 94.16%. This method effectively improves the economic efficiency and environmental friendliness of underground transportation, and provides a technical path for the construction of intelligent mines.