Spatial neglect (SN) impedes stroke rehabilitation progress, slows functional recovery, and increases caregiver stress and burden. The estimation of SN prevalence varies widely across studies. We aimed to establish the prevalence of SN based on the injured cerebral hemisphere, recovery stage post-stroke, and diagnostic methodology. All journal articles published up to February 27, 2019 from CINAHL, PsycINFO, PubMed and Web of Science were searched. We selected original research articles that described observational studies, included both individuals with left brain damage (LBD) and those with right brain damage (RBD) post-stroke, and reported specific diagnostic methods for SN. All authors reached consensus for the final selection of 41 articles. Time post-stroke, patient selection criteria, study setting, SN diagnostic methods were extracted. A total of 6324 participants were included: 3411 (54%) with RBD and 2913 (46%) with LBD. Without considering time post-stroke or diagnostic methods, the occurrence rate of SN was 29% (38% after RBD and 18% after LBD). Using ecological assessments resulted in higher prevalence than using tests not directly related to daily life activities (53% vs. 24%). Using methods based on a single-cutoff criterion led to lower occurrence of SN than using multi-test methods (27% vs. 33%). The prevalence decreased from the acute to chronic stage post-stroke. The estimated prevalence of SN after unilateral stroke is 30%. SN is more common after RBD than after LBD, but SN after LBD is still quite common. Using ecological assessments and multi-test methods to detect SN is preferred to using a single-cutoff criterion of a test that is not directly related to daily function. The decrease in SN prevalence over time is evident, but the exact prevalence in later stages cannot be estimated. More research is needed to better understand chronic SN.