Abstract: This study aims to establish a model for predicting the soluble solids content of tomatoes using electrical parameters and machine learning techniques. An LCR meter was employed to measure nine electrical parameters of tomatoes at frequencies of 0.1，1，10，100，and 1 000 kHz，including parallel equivalent capacitance，parallel equivalent resistance，and quality factor. Key electrical characteristic variables were identified through Pearson correlation analysis. Based on these variables，three non-destructive models for predicting the soluble solids content of tomatoes were constructed：a Back Propagation Neural Network（BpNN），Multiple Linear Regression（MLR），and Support Vector Regression （SVR）. The results indicate that at 10 kHz frequency，five electrical parameters–quality factor，loss factor，deflection angle，parallel equivalent capacitance，and parallel equivalent resistance–showed significant correlation with the soluble solids content of tomatoes. By using these five electrical parameters as input variables and the soluble solids content as the output variable，the SVR model exhibited the best predictive performance，coefficient of determination R2 equal to 0.951，Root Mean Square Error equal to 0.122，Mean Absolute Error equal to 0.082）. This study provides a new method for the rapid and non-destructive detection of soluble solids content in post-harvest tomatoes.

Key words: tomato, soluble solids content, electrical characteristics, machine learning, non-destructive testing