The association between 21 vegetables and eight fruits and prostate cancer risk was assessed in the Netherlands Cohort Study among 58,279 men of ages 55-69 years at baseline in 1986. After 6.3 years of follow-up, 610 cases with complete vegetable data and 642 cases with complete fruit data were available for analysis. In multivariate case-cohort analyses, the following rate ratios (RRs) and 95% confidence intervals (CIs) for vegetable consumption were found (comparing highest versus lowest quintile): total vegetables (RR, 0.80; CI, 0.57-1.12); prepared vegetables (RR, 0.85; CI, 0.61-1.19); and raw vegetables (RR, 0.96; CI, 0.69-1.34). For vegetables categorized in botanical groups, no associations were found except for consumption of pulses (RR, 0.71; CI, 0.51-0.98; P for trend, 0.01). The RRs for total fruit and citrus fruit were 1.31 (CI, 0.96-1.79) and 1.27 (CI, 0.93-1.73), respectively; the corresponding Ps for trend were 0.02 and 0.01, respectively. In a continuous model, no association for total fruit was observed. Individual vegetables and fruits were evaluated as continuous variables (g/day). Nonsignificant inverse associations (RRs per increment of 25 g/day) were found for consumption of kale (RR, 0.74), raw endive (RR, 0.72), mandarins (RR, 0.75), and raisins or other dried fruit (RR, 0.49). Observed positive associations were significant for consumption of leek (RR, 1.38) and oranges (RR, 1.07) and nonsignificant for sweet peppers (RR, 1.60) and mushrooms (RR, 1.49). Results in subgroups of cases were more or less consistent with the overall results. From our study, we cannot conclude that vegetable consumption is important in prostate cancer etiology, but for certain vegetables or fruits, an association cannot be excluded.