This study examines the dynamic impact of exchange rate volatility on non-performing loans (NPLs) in Iran’s banking system, while simultaneously accounting for the moderating role of bank-specific factors, including capital adequacy (CAR), liquidity (LQ), and management quality (MQ). A Panel Vector Autoregression (Panel-VAR) framework is employed using balanced panel data from 18 Iranian commercial banks over the period 2011–2023, comprising 234 bank-year observations. The adopted methodology integrates LLC and IPS unit root tests, Pedroni and Kao cointegration analyses, and system GMM estimation to address endogeneity and dynamic panel bias. Exchange rate volatility is measured using a GARCH(1,1) model based on the official USD/IRR exchange rate. The results indicate that exchange rate volatility exerts a positive and statistically significant effect on NPLs, operating primarily through the balance sheet channel. Capital adequacy and liquidity function as significant mitigating buffers, whereas management quality exhibits a smaller yet still significant negative effect. The high autoregressive coefficient of NPLs points to substantial persistence in the deterioration of credit quality. Cointegration analysis confirms the existence of three long-run equilibrium relationships among the variables, reinforcing the structural nature of these linkages. Granger causality tests establish bidirectional causality between exchange rate volatility and NPLs, suggesting the presence of reinforcing feedback dynamics. The findings support prioritizing exchange rate stabilization as a prerequisite for financial stability, accelerating the implementation of Basel III capital requirements with countercyclical buffers calibrated to exchange rate risk, and integrating exchange rate indicators into banking supervisory early warning systems.