<p>In post-conflict cities, ethnic segregation is an engineered process driving spatial inequality. Using Kabul, Afghanistan, as a critical case study, this research develops a conflict-sensitive analytical framework to explicate the causal pathways linking politicized segregation to disparities in access to public services, with a primary focus on healthcare. We employ a sequential explanatory mixed-methods design. To address the challenge of modeling non-physical barriers in a conflict zone, we introduce a Security-Risk Weighted Two-Step Floating Catchment Area (2SFCA) model that integrates ACLED conflict data with systematically coded and validated ethnographic narratives. We also construct a Politicized Segregation Index (PSI), grounded in a Delphi expert elicitation process, to move beyond purely demographic measures. These novel metrics are then integrated into Spatial Autoregressive (SAR) and Structural Equation Modeling (SEM) analyses. Results confirm a significant negative impact of segregation on accessibility (β = –0.538, p &lt; 0.001), with strong spatial dependence (ρ = 0.412, p &lt; 0.001). SEM further shows that this relationship is fully mediated by historical disinvestment (38%) and security-related mobility barriers (25%). Qualitative interviews (n = 50) corroborate that ethno-political patronage drives infrastructure allocation, while checkpoints and fear diminish the functional value of physical proximity. This integrative analysis provides a mechanistic explanation of urban inequality, yielding a replicable framework for fragile cities and a compelling case for equity-driven planning that re-centers connectivity and mobility security.</p>