The challenges in vehicle pickup and distribution in the automobile industry regarding efficiency, environmental issues, and satisfaction are gradually rising due to complex same pickup and delivery issues in fixed timelines. The current research contributes to the development of an effective model using machine learning algorithms [18] to optimize vehicle routing in real-world scenarios from collection points, seaports, to final distribution points. The scenario model problem was formed as multi-objectives including (1) minimizing costs, (2) reducing carbon emissions, and (3) maximizing satisfaction related to schedule following. To proceed with the model process, an enhanced Genetic Algorithm model named EGA was developed with dynamic adaptation strategies coupled with K-means++ and delay prediction using the random forest algorithm. In applying the developed model using real data from 250 actual cases, there was an increase in operational accuracy of 88.89%, a reduction cost of 22.6%, reduction in carbon emissions per month of 19.3 tons, in addition to boosting successful deliveries from the current 76% to 92% in the actual cases. Using the NSGA-II approach, an optimal solution in terms of Pareto front was derived in applying the developed model by highlighting proper trade-offs under economic, carbon, and time factors. An optimal strategic balance was obtained using the current research by applying the developed model; hence, theoretically, the current research developed and expanded the VRP model framework to involve same deliveries/receipt, variable timeline considerations, and carbon emissions being the prime factor. In terms of real-world applications, the current research contributes towards developing smart logistics computerized decision-making support related to sustainable green logistics.