The humidity within the leaf boundary layer significantly deviates from the surrounding humidity, plant diseases are often closely related to humidity. Predicting leaf boundary layer humidity is important for preventing plant diseases. Consequently, grapes ("Beibinghong", "Jufeng") and strawberries ("Hongyan", "Ningyu") were chosen for thorough research analysis. First, a comprehensive investigation was carried out to examine the trends of leaf boundary layer humidity under different weather conditions and soil moisture treatments. Second, an in-depth analysis was conducted to clarify the correlations between leaf physiological indices and leaf boundary layer humidity. Employing the principles of Support Vector Regression (SVR), a predictive model for leaf boundary layer humidity was then developed. The results showed that: (1) plant leaf boundary layer humidities are significantly higher than ambient humidity at distances of 1 mm and 5 mm from both the upper and lower leaf surfaces, with no significant difference observed at distance of 15 mm. In addition, the highest leaf boundary layer humidity is found at distance of 1 mm from the upper and lower leaf surfaces. Moreover, the discrepancy between leaf boundary layer humidity and ambient humidity was more conspicuous on sunny days than cloud days; (2) under both sunny and cloudy conditions, the humidities within the leaf boundary layer are consistently higher on the lower surfaces of leaves compared to their upper surfaces. Additionally, as soil moisture content increases, the humidities at distances of 1 mm and 5 mm from the leaf surfaces exhibit a corresponding elevation; (3) the humidities within the leaf boundary layer demonstrate a highly striking and positive correlation with net photosynthetic rate (P_n), transpiration rate (T_r), leaf water potential, stomatal length (SL), and soil moisture content. Nonetheless, the above humidities display a highly remarkable and negative correlation with ambient humidity and distance from the upper and lower leaf surfaces; the indicators are ranked based on their correlation with the leaf boundary layer humidity as follows: the distance between the upper and lower leaf surfaces, environmental humidity, net photosynthetic rate, soil moisture content, leaf water potential, stomatal length, and transpiration rate. (4) the leaf boundary layer humidity prediction model based on the Support Vector Regression (SVR) had a coefficient of determination R² of 0.938, which is above 0.9, which clearly illustrates a desirable fit and superior precision. The leaf boundary layer humidity prediction model allows for rapid and precise forecasting of leaf boundary layer humidity, which is important for ecological control of diseases and provides a theoretical basis for studying the relationship between crop cultivation and the environment.