Quality of finished Cut-Tear-Curl (CTC) tea mainly depends on biochemical components like Thearubigin (TR) and Theaflavin (TF). Traditional estimation of TF and TR requires analytical instruments. These are expensive, and require long time and laborious effort to determine TF and TR. This paper presents an effective method to estimate the content of TF and TR of tea samples using Electronic Tongue (ET) response. A regression model is developed using the features extracted from the ET signals to predict TF and TR content of tea. Energy values of the signals from different sensors of ET computed by the coefficients of Walsh Hadamard Transform (WHT) are used as feature to develop regression models. Two different models such as Support Vector Regression (SVR) and Vector-Valued Regularized Kernel Function Approximation (VVRKFA) are used to justify the performance of the proposed method. High prediction accuracy ensures the usefulness of the proposed method for the prediction of TF and TR using ET signals.