This paper proposes a practical and transparent framework for determining recommended vehicle speeds under varying pavement and environmental conditions. The method links a road-condition classification (dry-smooth, dry-rough, wet, icy) to an explicit engineering decision rule based on two safety constraints: (i) a stopping-distance requirement and (ii) a curve-stability requirement for representative horizontal-curve scenarios. When continuous field measurements are unavailable, friction coefficients are adopted as typical literature- and standards-based ranges for scenario calculations, while the resulting speeds are interpreted as indicative values subject to local calibration. To quantify uncertainty caused by variability in tire-road friction and driver reaction time, a Monte Carlo simulation is implemented with uniform sampling of friction within class ranges and reaction time within 0.8-1.2 s (Nsim = 10,000). The output is reported as the median recommended speed and the 95% uncertainty interval [P2.5, P97.5] for each pavement class, enabling conservative engineering margins. Under the studied assumptions (Vlimit = 110 km/h, D = 80 m, ksafety = 0.90), the median recommended speed decreases from 92.0 km/h for a representative dry-curve case to 71.2 km/h for wet straight segments and 45.2 km/h for an icy curve. The proposed approach supports condition-responsive speed management and can serve as a computational core for future ITS deployment after calibration using local friction and surface-condition measurements.