Leaf area (LA) is a fundamental growth variable for understanding physiological processes and crop yield, but direct measurements are often destructive and time-consuming. This study aimed to develop non-destructive allometric models for estimating the leaf area of sweet potato (Ipomoea batatas [L.] Lam.) cultivars using leaf dimensions. Morphological data, including leaf length (L), width (W), and their product (L × W), were collected from various cultivars. Linear, power, and exponential models were tested and validated using statistical indices such as the coefficient of determination (R²), root mean square error (RMSE), and Willmott’s d-index. The results indicated that models based on the product of length and width (L × W) provided the most accurate estimations within the evaluated dataset, exhibiting high predictive accuracy (R² ≥ 0.96, RMSE ranging from 4.24 to 5.19 cm², and Willmott’s d-index ≥ 0.98). This non-destructive approach can support rapid, precise leaf area monitoring without leaf harvesting, particularly in the evaluated cultivars and experimental conditions. In conclusion, the proposed allometric equations provide practical and accurate tools for estimating sweet potato leaf area and supporting growth assessments in similar experimental contexts.