For the collaborative transportation demand between Ro-Ro trailer terminals and inland customers with multiple task types, a tractor scheduling optimization model with time windows, trailer capacity constraints, and considering the inherent correlation between task types and node access order was constructed, and a mixed integer programming model with the goal of minimizing total  transportation costs was established. An improved adaptive large neighborhood search algorithm was proposed to enhance the search efficiency of the destroy and repair operators through a two-stage feasible arc generation strategy and a dynamic weight update mechanism. Numerical experiments demonstrate the correctness of the model and the effectiveness of the algorithm. The experimental results show that the flexible node visit strategy can coordinate the tractor and trailer resources without violating time and capacity constraints, reducing the total transportation cost by 3.16% to 6.47%. Compared with the traditional separation mode, the Ro-Ro separation mode can reduce the total cost by 5.29% to 12.71%,  decreasing the number of tractors required for operation by improving resource utilization. This study can provide a decision-making reference for resource coordination and scheduling between Ro-Ro terminals and the hinterland.