Questions What are the degree and pattern of dual-task interference during walking in people after stroke? How do these vary with disease chronicity and different component tasks in people after stroke? How does dual-task interference differ between people after stroke and people without stroke? Design Systematic review with meta-analysis of studies reporting gait-related dual-task interference. Participants People after stroke and people without stroke. Outcome measures Measures of walking and secondary (cognitive or manual) task performance under dual-task conditions relative to those under single-task conditions. Results Seventy-six studies (2,425 people after stroke and 492 people without stroke) were included. Manual and mental tracking tasks imposed the greatest dual-task interference on gait speed, although there was substantial uncertainty in these estimates. Among mental tracking tasks, the apparently least-complex task (serial 1 subtractions) induced the greatest dual-task interference (−0.17 m/s, 95% CI −0.24 to −0.10) on gait speed, although there was substantial uncertainty in these estimates. Mutual interference (decrement in both walking and secondary component task performances during dual-tasking) was the most common dual-task interference pattern. The results of the sensitivity analyses for studies involving people with chronic stroke were similar to the results of the primary analyses. The amount of dual-task interference from a mental tracking or manual task during walking was similar between people with or without stroke. Conclusions The degree and pattern of dual-task interference vary with the choice of component tasks. When evaluating limitations to functional mobility during dual-tasking conditions and in planning interventions accordingly, clinicians should select dual-task assessments that correspond to the daily habits and physical demands of people after stroke. Registration CRD42017059004.