Abstract Background Plant-based dietary patterns are increasingly popular in western countries and are supported by many governments and health organisations for their potential beneficial role in the prevention of chronic diseases. Yet, the potential role of plant-based dietary patterns in the development of dementia remains unclear. Objective To evaluate the association between plant-based dietary patterns and the risk of dementia. Methods Dietary intake was measured at baseline in 9,543 dementia-free participants (mean age 64 years, birth years 1897–1960, 58% women) of the prospective population-based Rotterdam Study, using food frequency questionnaires. Based on these questionnaires, we calculated an overall plant-based dietary index (PDI), healthy PDI (hPDI) and unhealthy PDI (uPDI), with higher scores reflecting higher consumption of (any, healthy and unhealthy, respectively) plant-based foods and lower consumption of animal-based foods. We analysed the association of the PDIs with incident dementia using Cox proportional hazard models. Results During a mean follow-up of 14.5 years, 1,472 participants developed dementia. Overall, the PDIs were not associated with the risk of dementia (hazard ratio [95% confidence interval] per 10-point increase: 0.99 [0.91–1.08] for PDI, 0.93 [0.86–1.01] for hPDI, 1.02 [0.94–1.10] for uPDI). However, among men and APOE ε4 carriers, a higher hPDI was linearly associated with a lower risk of dementia (0.86 [0.75–0.99] and 0.83 [0.73–0.95], respectively), while this association was U-shaped among APOE ε4 non-carriers (P value for non-linearity = 0.01). Conclusions We found no strong evidence for an overall association between plant-based eating and the risk of dementia. Our findings in stratified analyses warranted further investigation.