In this study, we aimed to develop a mixed tea prepared with roasted mulberry leaf and fruit using response surface methodology (RSM). Roasting of mulberry leaf was by 6 stages, as shown in Fig. 1; and mulberry fruit was roasted in 4 stages, as shown in Fig. 2. Subsequently, physicochemical measurements such as total polyphenol content, nitric oxide production content, and α-glucosidase inhibitory effect were obtained for each sample. Central composite design was applied to prepare samples containing varying contents of roasted mulberry leaf (RoML) and roasted mulberry fruit (RoMF); subsequently, sensory evaluation was conducted. The total polyphenol content of roasted samples (RoML and RoMF) were significantly higher than that of raw samples (RaML and RaMF), respectively. The nitric oxide (NO) production of roasted samples were significantly lower than that of control (LPS induced RAW 264.7 cell). The α-glucosidase inhibitory effect of roasted samples was significantly higher than that of raw samples, respectively. Based on the RSM estimation for determination of optimum ratio by sensory evaluation (taste, color, and flavor) among 13 mixed samples, the optimum mixing ratio of RoML and RoMF for taste, color, and flavor were 1.64 g (RoML) and 0.88 g (RoMF), 1.35 g (RoML) and 0.92 g (RoMF), 1.65 g (RoML) 1.03 g (RoMF), respectively. Based on results of three sensory evaluations, mixing ratio comprising 1.54 g of RoML and 0.92 g of RoMF is desirable for delicious tea with functionality.