This study aims to quantitatively verify the noise reduction effectiveness of variable speed limit (VSL) control strategies prior to the development of an active road traffic noise control system. Using a microscopic traffic noise (MTN) model, real-time time-series noise levels were calculated for each speed control scenario, and the resulting noise reductions were compared. The MTN model treats each individual vehicle moving according to a car-following model as a moving noise source, and was implemented using the open-source microscopic traffic simulator SUMO. Single-vehicle experiments demonstrated that reducing the speed limit decreased the maximum noise level at a receiver 10 m from the road edge by 4.67–15.79 dB(A) for passenger cars and 0.59–13.45 dB(A) for heavy vehicles. Spatial analysis revealed that noise reduction for passenger cars decreased monotonically with receiver distance. For heavy vehicles, the maximum noise reduction occurs at an offset distance of 20–100 m from the road edge rather than at the nearest point. In the multivehicle experiment, a speed reduction from 100 to 80 km/h yielded an equivalent noise level reduction of up to 1.59 dB(A). These results, which cannot be captured by conventional macroscopic noise prediction models, demonstrate the feasibility of MTN-based active noise control and provide a quantitative basis for the design of VSL algorithms aimed at road traffic noise reduction.