In recent years, Iran’s foreign exchange market has experienced sharp exchange-rate jumps, heightened volatility, and rapid shifts in economic agents’ expectations—phenomena that conventional exchange-rate models have struggled to explain. This study develops an agent-based model comprising technical and fundamentalist traders. The model includes 1,000 heterogeneous agents who choose trading strategies probabilistically, with probabilities depending on the deviation of the exchange rate from their perceived fundamental value. Exchange-rate dynamics are also affected by heterogeneous beliefs, trading shocks, inventory control, the long-run trend, sanctions-related developments, and central bank responses.Model parameters are calibrated using the Method of Simulated Moments (MSM) and a genetic algorithm, based on 4,258 daily log returns from 1390 through the end of Azar 1404 in the Solar Hijri calendar. The objective function includes 18 moments capturing return-distribution shape, tail percentiles, temporal dependence, and volatility clustering.Results from 500 independent Monte Carlo runs show that 16 of the 18 key return moments fall within the 90% simulation interval. The model satisfactorily reproduces return mean and dispersion, high kurtosis, fat tails, and volatility persistence, although it understates negative return autocorrelation at some very short lags.The findings indicate that exchange-rate instability in Iran is not solely driven by external shocks such as economic sanctions. Nonlinear interactions between heterogeneous expectations and trading strategies also generate endogenous volatility. This underscores the importance of jointly managing external shocks and endogenous market mechanisms by reducing uncertainty, enhancing information transparency, and implementing targeted interventions.