The present work utilizes the molecular imprinted polymer technology for the fabrication of a stable and reproducible electrode for the detection of epicatechin (EC) in tea. The polymer-graphite composite embedded with amine functionalized multi-walled carbon nanotubes (f-MWCNTs) was prepared and used as the sensing material. The sensing materials of the MIP-EC electrode were characterized by field emission gun scanning electron microscopy, transmission electron microscopy and fourier transform infrared spectroscopy. A wide linearity change from <inline-formula> <tex-math notation="LaTeX">$1~ \boldsymbol {\mu }\text{M}$ </tex-math></inline-formula>– <inline-formula> <tex-math notation="LaTeX">$30~ \boldsymbol {\mu }\text{M}$ </tex-math></inline-formula> and 30 <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {\mu }\text{M}$ </tex-math></inline-formula> - 300 <inline-formula> <tex-math notation="LaTeX">$\boldsymbol {\mu }\text{M}$ </tex-math></inline-formula>, detection limit of 51.92 nM and quantification limit of <inline-formula> <tex-math notation="LaTeX">$0.17~ \boldsymbol {\mu } \text{M}$ </tex-math></inline-formula> were obtained with the electrode. To predict the content of EC, two linear regression models were developed, using the differential pulse voltammetry responses and high-performance liquid chromatography analysis results as reference data. Prediction accuracies of 92% using principal component regression (PCR) and 93.58% using partial least square regression (PLSR) were obtained. The root mean square error of calibration for PLSR and PCR was 0.29 and 0.31 using 6 and 10 latent variables, respectively. The stability of the electrode was found to be of approximately 90 days.