This study investigates the effects of changes in the cycling environment on cyclists’ dynamic behavioral responses and explores how the environment contributes to the cyclists’ overall subjective rating of safety and comfort. Twenty-four participants cycling through virtual-reality simulations of environments of a local town road in Japan were exposed to six infrastructure design scenarios having different widths, bike lane colors, and lane separation types in the virtual reality. The estimation results of one-boundary drift-diffusion decision models across the different scenarios showed that designs with clear separations from car lanes facilitated faster reactions, aligning with higher subjective safety ratings. Moreover, a changepoint analysis was conducted to assess the cycling performance in terms of braking and steering over the different designs. The main conclusion of this study is that a drift-diffusion decision model can be applied to the analysis of cycling behaviors and such analysis might provide insights into safety infrastructure evaluation.