Memory plays a central role in food choice and consumption. Most recent studies focusing on food memory and its role in everyday eating and drinking behaviour used a paradigm developed by Mojet and Köster (2002), based on incidental learning of target foods, and an unexpected memory test, demanding recognition of the target. The present work was an exploratory research, and the objective was to study the impact of the retention interval between the learning phase and the recognition phase. Three groups of participants were offered a target orange juice during a first meal organised under a false pretence to ensure incidental learning. Depending on the groups, participants came back 1 h, 1 day or 1 month later to perform a recognition task: They had to recognise the target orange juice amongst orange juices that differed slightly from the target on sweetness or bitterness. Results showed a rapid decline in recognition performance as the retention interval increased in length. This deterioration over time seems to show that incidentally learned memory is not as resistant to decay as intentionally learned olfactory memory.