Traffic management in a smart city is important in several aspects. From wasting people's time in traffic, which causes financial and mental harm to them, to increasing air pollution, which can directly affect their physical health and have destructive effects on the environment. Providing a mathematically based system in which drivers can choose the optimal driving time, drive at the highest speed and the lowest pollution on the street, can be a good help to traffic management in a smart city. The presented Modified Best-Mean-Random (MBMR) algorithm is a population-based metaheuristic designed to optimize multi-criteria traffic parameters (e.g., traffic volume, speed, pollution) on Stadhouderskade street in Amsterdam. The performance of the proposed MBMR algorithm was benchmarked against three competitive metaheuristics—GWO, MOEA/D, and BES—based on accuracy, stability, and computational efficiency. MBMR achieved the highest accuracy with a 39.02%, 21.88%, and 34.21% improvement over GWO, MOEA/D, and BES, respectively. In terms of stability, it consistently outperformed all other algorithms, showing a 55.56% to 63.64% reduction in standard deviation across multiple runs. While its execution time was marginally higher than GWO by 2.4 seconds, it remained notably faster than MOEA/D and BES. These results confirm MBMR's capability to deliver precise, stable, and computationally efficient solutions, making it a robust candidate for real-world multi-objective optimization tasks.