Therapeutic radiotherapy is an important treatment of pelvic cancers. Historically, low-fiber diets have been recommended despite a lack of evidence and potentially beneficial mechanisms of fiber. This randomized controlled trial compared low-, habitual-, and high-fiber diets for the prevention of gastrointestinal toxicity in patients undergoing pelvic radiotherapy. Patients were randomly assigned to low-fiber [≤10 g nonstarch polysaccharide (NSP)/d], habitual-fiber (control), or high-fiber (≥18 g NSP/d) diets and received individualized counseling at the start of radiotherapy to achieve these targets. The primary endpoint was the difference between groups in the change in the Inflammatory Bowel Disease Questionnaire-Bowel Subset (IBDQ-B) score between the starting and nadir (worst) score during treatment. Other measures included macronutrient intake, stool diaries, and fecal short-chain fatty acid concentrations. Patients were randomly assigned to low-fiber ( = 55), habitual-fiber ( = 55), or high-fiber ( = 56) dietary advice. Fiber intakes were significantly different between groups ( < 0.001). The difference between groups in the change in IBDQ-B scores between the start and nadir was not significant ( = 0.093). However, the change in score between the start and end of radiotherapy was smaller in the high-fiber group (mean ± SD: -3.7 ± 12.8) than in the habitual-fiber group (-10.8 ± 13.5; = 0.011). At 1-y postradiotherapy ( = 126) the difference in IBDQ-B scores between the high-fiber (+0.1 ± 14.5) and the habitual-fiber (-8.4 ± 13.3) groups was significant ( = 0.004). No significant differences were observed in stool frequency or form or in short-chain fatty acid concentrations. Significant reductions in energy, protein, and fat intake occurred in the low- and habitual-fiber groups only. Dietary advice to follow a high-fiber diet during pelvic radiotherapy resulted in reduced gastrointestinal toxicity both acutely and at 1 y compared with habitual-fiber intake. Restrictive, non-evidence-based advice to reduce fiber intake in this setting should be abandoned. This trial was registered at clinicaltrials.gov as NCT 01170299.