This study investigates the relationship between inflation and short‐term exchange rate movements in three Western Balkan economies, North Macedonia, Albania, and Serbia, each operating under a distinct monetary regime. We employ monthly data spanning January 2018 through December 2023 and apply decision tree classifiers to determine whether inflation alone can predict whether a currency will appreciate, depreciate, or remain stable. In the fixed regime of North Macedonia (denar pegged to the euro) and the managed float of Serbia (dinar), inflation has virtually no predictive power, as central bank interventions maintain prolonged stability. In contrast, Albania’s fully floating lek exhibits moderate sensitivity: higher inflation is associated with depreciation episodes, though appreciations remain unpredictable. Rolling six‐month correlations between inflation and exchange rate changes confirm these findings, revealing only transient co‐movements during major shocks. Our results suggest that in rigid regimes, central banks effectively neutralize inflationary pressures, whereas in floating settings, inflation can signal currency weakness. Nevertheless, a single‐feature machine learning model struggles with class imbalance, accurately classifying stability but failing to forecast appreciations or depreciations with confidence. We conclude by recommending richer feature sets and more sophisticated models for future research and by discussing policy implications for emerging‐market analysis.