In traffic safety research, exposure indicators quantify the conditions and opportunities under which accidents may occur. To overcome the limitation that conventional maritime traffic exposure indicators fail to capture the variability of the navigational environment or the differences arising from vessel characteristics, this study proposes the Navigational Constraint Index (NCI), an index framework that quantifies navigational constraint from vessel–environment interaction. Dividing the Korean coastal waters into approximately 5 km × 5 km grid cells, the framework operationalizes the two physical dimensions through which constraint acts as distinct safety concepts. The Horizontal Navigational Constraint Index (H-NCI), defined as the ratio of the cumulative ship-domain occupancy area (scaled to vessel length) to the effective navigable area, is an exposure indicator that expresses the density of accident opportunities per unit grid. The Vertical Navigational Constraint Index (V-NCI), defined as the ratio of the required depth (static draft combined with dynamic allowances for squat, heel, and wave-induced motion and a minimum under-keel clearance) to the available water depth, is a vessel-specific risk indicator expressing proximity to a grounding condition. H-NCI is validated against four criteria for qualifying as an exposure measure (construct validity, concurrent evidence, exposure-intensity gradient, and discriminant validity); to this end, Geographically Weighted Poisson Regression (GWPR) and the machine-learning-based PyGALAX framework are applied to concurrently identify the spatial heterogeneity of accident-influencing factors. V-NCI, by contrast, is derived deterministically from a physical limit-state relationship; applied to six representative vessel-type and draft conditions configured from port arrival and departure records, it assesses how the spatial distribution of vertical risk shifts with increasing draft, while a sensitivity analysis confirms the stability of its critical-grade classification. By proposing a constraint-based spatial indicator framework, this study provides an analytical basis for spatially prioritized decision-making in maritime safety management across both exposure and risk dimensions.