Chemotherapy preparations are often complex and subject to a strict regulatory context. The existing control methods are often limited to Double Visual Control (DVC). In this paper, the preparation circuit of chemotherapy drugs is evaluated through data collection and statistical analysis in order to highlight the difficulties encountered. The results regarding preparation and control times and the number of task interruptions highlight the unreliability of the DVC and its impact on processing time. As a solution, we propose a decision support system “Smart Prep” based on Augmented Reality (AR), co-developed, and commercialized by the Faculty of Pharmacy of Lille, Ecole Centrale de Lille and the company Computer Engineering. This system allows the preparation of chemotherapy drugs according to a step-by-step mode, a traceability of the preparation steps and a reduction of tasks’ interruptions.