Criminal behavior has been traditionally linked to green space. However, studies on the relationship between green space and crime vary in their conclusions and even contradict each other. This variability is primarily attributed to the mixing-up of specific crimes and the less-explored influencing mechanisms. This study considers Toronto and focuses on a specific kind of crime, assault, for exploring whether green space influences the neighborhood assault rate. If so, then how does it affect the assault rate? By considering the neighborhood as the spatial analysis unit, the indicators of green space and its component structure (namely the composition of trees, shrubs, and grass) are included in the regression analysis model used in this study. The dual effect (both positive and negative) of walkability on crime is treated as the moderator in the regression model. Our analysis revealed that high percentages of green space, tree area, and grass area potentially inhibit assault occurrence, while the shrub area percentage has no statistically significant association with the assault rate. When walkability reaches a certain degree (84th or/and 50th percentiles in this study), it can enhance the inhibitory impact of the green space and its component structure (including trees, shrubs, and grass) on the assault rate. Comparatively, this reduction effect is more significant for green space composed of grass. Therefore, we can infer that increasing the percentage of trees and grass, while simultaneously improving walkability, can reduce residents’ risk of being assaulted at the neighborhood level.