White light-emitting diodes (WLEDs) for road lighting are required to have both the high scotopic to photopic ratio (S/P) and color rendering index (CRI). However, there is a trade-off between S/P and CRI, and WLEDs commonly having the S/P of 1.68-2.38 usually exhibit a low CRI. In this work, to provide a best solution to the trade-off problem we proposed an optimal spectral model for the phosphor-converted WLEDs (pc-WLEDs) aiming to figure out the optimal phosphor combination. The Monte Carlo Algorithms combined with the Genetic Algorithm were adopted to obtain the optimization of CRI and S/P by varying the spectral power distributions of WLEDs through adjusting the spectral parameters. Considering the spectral requirements of pc-WLEDs based on the mesopic vision, we chose CaAlSiN 3 :Eu 2+ and Y 3 (Ga,Al) 5 O 12 :Ce 3+ as the red-and green-emitting phosphors to prepare WLEDs with both high S/P and CRI, respectively. The simulation based on the optimal spectral model led to an optimal pc-WLED with a high S/P of 2.0-2.14, Ra > 80 and correlated color temperature (CCT) of 4000-5000 K, which matches very well with the experimental results of S/P = 2.064, Ra = 93.9, and CCT = 4981 K for the two-phosphor converted WLEDs. It implies that the optimal spectral model would be used effectively for the spectral design and phosphor selection for WLEDs.