Abstract: Tannins are the most abundant phenolic compounds after cellulose, hemicellulose, and lignin. Condensed tannins are one class of tannins with diverse functional activities, such as hemostatic and wound healing, anti-inflammatory and anti-allergic behavior. There is a key research topic in the field of plant polyphenols. Banana peel is one of the main by-products of the banana processing industry, which is usually used as animal feed or directly landfilled with a low utilization rate of the functional components. In fact, banana peel can contain many active ingredients, such as condensed tannins. Previous research has been widely used to optimize the extraction parameters for condensed tannins of banana peel. The structural analysis has been conducted to demonstrate their antioxidant, inhibition of Escherichia coli, and anti-hyperglycemic effects. However, the extracted condensed tannins of banana peel are required for purification. Few purification studies have been carried out at present. In this study, macroporous resins were used to purify the condensed tannins of banana peel. A comparison was made on the static adsorption and desorption performance of six macroporous resins. The most suitable resin was selected in this case. Kinetic models, adsorption isotherm, and thermodynamic parameters were then used to analyze the adsorption properties of macroporous resins. A systematic investigation was implemented to clarify the effects of loading concentration, loading flow rate, and volume on adsorption efficiency, as well as the effects of ethanol concentration, elution flow rate, and volume on desorption efficiency. The results showed that the HP-20 macroporous resin shared a high adsorption rate (87.47%) and desorption rate (84.12%), indicating the best purification for the condensed tannins of banana peel. The adsorption kinetics curve demonstrated that the adsorption capacity of tannin on HP-20 resin sharply increased within the first 4 h, thus reaching the maximum adsorption capacity in 8 h, and then maintaining the stability. The desorption kinetics proved that the desorption was essentially realized within 5 h, with a desorption rate of 92.87%. The adsorption curve was fitted to the pseudo-second-order kinetics (R²=0.997 2). Additionally, the adsorption isotherm experiment confirmed that the adsorption of banana peel condensed tannins by HP-20 resin increased, as the temperature raised. Therefore, the higher temperature was favored to the adsorption. The adsorption isotherms curves were fitted using thermodynamic equations. Freundlich thermodynamic model was suitable for the adsorption behavior of banana peel condensed tannins on HP-20, indicating multilayer sorption on a heterogeneous surface. The K_F of the Freundlich model increased with the increasing temperature, indicating that the higher temperature was favorable for the adsorption. In both the Langmuir and Temkin models, the parameters R_L and K_T were used to predict the adsorption from different perspectives. The computed value of R_L was less than 1, indicating the favorable adsorption of banana peel condensed tannins onto HP-20 resins. Additionally, K_T increased with temperature rising. Consequently, there was a stronger binding energy between condensed tannins and the resin, resulting in a tighter association between them, as temperature increased. Moreover, the thermodynamic parameter was ΔG<0, ΔH=5.13 kJ/mol, ΔS=47.28 J/mol·K. Therefore, the adsorption process was spontaneous and endothermic, with an increased randomness and dominant physisorption. The optimal conditions for purification were determined as the sample concentration of 6 mg/mL, ethanol concentration of 70%, and loading flow rate of 1.5 mL/min, for loading and elution, using volumes of 200 and 120 mL. The purity of banana peel condensed tannins increased from 5.79% to 68.44%. The HP-20 resin and purification conditions were effectively improved for the purity of banana peel condensed tannins.