Temporal Weighting Rule (TWR) is a mathematical model used to predict foraging decisions based firstly on recency and secondly on the quality of a patch (amount of food). The objective of this study was to investigate whether those variables modulate the decisions of spider monkeys (Ateles geoffroyi) using a foraging task. Sixteen adult individuals completed a series of three experiments where they had to recover a reward from one of three containers that simulated foraging patches. The first experiment showed that spider monkeys prefer higher quality patches over recently visited patches. The second experiment showed that spider monkeys take the total amount of reward into account, not the amount of reward per trial. Finally, in the third experiment, the data showed that spider monkeys do not choose patches based on effort, but on quality. We concluded that spider monkeys choose a patch based firstly on quality, and secondly on recency, contrary to what is proposed by the TWR. Additionally, the quality values of the patches are stored globally, and they remember the quality and location of patches for periods as long as 24 hrs, which implies the involvement of a long-term memory process.