Abstract: Optimize the formula process of Dendrobium officinale granules using a multi index comprehensive scoring method combined with orthogonal experiments. Based on the registered Dendrobium officinale health food, the formula composition of raw materials in Dendrobium officinale health food was determined by using complex system entropy clustering algorithm. The diluent was filtered by the formation rate, dissolution rate and moisture absorption rate of granules. The ratio of raw materials and diluents, ethanol volume concentration, and ethanol dosage were taken as the inspection factors. The analytic hierarchy process (AHP) and entropy weight method (EWM) were combined to establish a multi-index comprehensive evaluation method (AHP-EWM method). The formulation process of granules was optimized through orthogonal experiments. The results showed that the formula combination was Dendrobium officinale, Lycium chinense Mill., and Paeoniae Radix Alba by the complex system entropy clustering algorithm. The AHP-EWM method determines the weights of each indicator as follows: Molding rate 0.3567, angle of rest 0.0847, melting rate 0.1778, and moisture absorption rate 0.3808. The optimal molding process was the raw material extract powder, dextrin, and microcrystalline cellulose (10:7:3, m/m), 50% ethanol volume concentration, and 5 mL ethanol dosage. The valieation test result showed that the AHP-EWM comprehensive score of the granules was 93.63±1.65, which was close to the theoretical prediction value. The optimized granules formulation process was stable and feasible. This study would provide a reference for the development of health products of Dendrobium officinale.