The study aims to evaluate the potential of hyperspectral leaf reflectance data for estimating potassium (K) content in two vegetable crops—sugar beet (Beta vulgaris L., cv. Tapir) and celery (Apium graveolens L., cv. Neon). The research seeks to determine whether this approach can provide a reliable, non-invasive, and cost-effective method for monitoring K levels across different plant species, K concentration ranges, and developmental stages, thereby supporting precision fertilization management. Timely and precise monitoring of nutrient levels in plant foliage is crucial for optimizing fertilizer management, enhancing crop productivity, and ensuring environmental sustainability. Spectral reflectance in the visible and near-infrared ranges was recorded across a wide spectrum of K fertilization levels and three plant developmental stages. A Random Forest (RF) regression model was applied to identify optimal wavelengths for predicting foliar K content, using both species-specific and combined datasets. The RF model using 12 selected wavelengths (425, 443, 479, 599, 631, 662, 798, 863, 897, 921, 1978, and 2053 nm) achieved high prediction accuracy for K content in sugar beet (R² = 0.85), celery (R² = 0.79), and both species combined (R² = 0.81). The study identified key spectral features that correspond with physiological changes related to anthocyanin, carotenoid, chlorophyll b, starch, and protein contents. Spectral reflectance offers a reliable, non-invasive, and cost-effective method for estimating K content in plant leaves. The findings support its potential application in precision fertilization strategies across different vegetable species and growth stages.