Objectives Antipsychotics are widely used to treat first-episode psychosis but may have an anticholinergic burden, that is, a cumulative effect of medications that block the cholinergic system. Studies suggest that a high anticholinergic burden negatively affects memory in psychosis, where cognitive deficits, particularly those in verbal memory, are a core feature of the disease. The present study sought to replicate this in a large cohort of well-characterized first-episode psychosis patients. We expected that patients in the highest anticholinergic burden group would exhibit the poorest verbal memory compared to those with low anticholinergic burden and healthy controls at baseline (3 months following admission). We further hypothesized that over time, at month 12, patients’ verbal memory performance would improve but would remain inferior to controls. Methods Patients ( n = 311; low anticholinergic burden [ n = 241] and high anticholinergic burden [ n = 70], defined by a Drug Burden Index cut-off of 1) and healthy controls ( n = 128) completed a clinical and neurocognitive battery including parts of the Wechsler Memory Scale at months 3 and 12. Results Cross-sectionally, using an analysis of variance, patients in the highest anticholinergic burden group had the poorest performance in verbal memory when compared to the other groups at month 3, F(2,430) = 52.33, P < 0.001. Longitudinally, using a Generalized Estimating Equation model, the verbal memory performance of all groups improved over time. However, patients’ performance overall remained poorer than the controls. Conclusion These findings highlight the importance of considering the anticholinergic burden when prescribing medications in the early stages of the disease.