Drug planning is the process of activities in the selection of types, quantities, and prices in accordance with the needs and budget for the coming procurement period in order to avoid the occurrence of excess or emptiness of drug supplies when needed by patients. Management of planning that is not optimal drug needs will have a negative impact on hospitals, both medically and economically, because 50-60% of the total budget used for treatment and medical equipment, uncertainty of drug needs due to disease population and the number of patients can change according to conditions the volume of patient diagnostic data, thus requiring an automatic way to select drug needs according to disease progression. This study aims to create a prediction model for drug needs with the ANFIS method, the data analysis framework used is sourced from drug usage / sales data at the Royal Prima Hospital 2016-2017 by building a software that implements the ANFIS method. Stages of application testing are carried out by applying the previous year's data to predict the current year, namely the 2016 data for 2017 predictions, while the 2017 data for 2018 predictions. The data source will be used to analyze the ANFIS membership function that generates parameters for the ANFIS method in training and testing data. The results of the analysis of the ANFIS parameters will be updated to produce a small error value (close to 0), based on the value of Root Mean Square Error (RMSE), then an evaluation is carried out with a quantitative and qualitative analysis of the predicted results with the existing system.