BACKGROUND: Homicide is an important cause of premature mortality globally, but evidence for the magnitude of homicides by intimate partners is scarce and hampered by the large amount of missing information about the victim-offender relationship. The objective of the study was to estimate global and regional prevalence of intimate partner homicide. METHODS: A systematic search of five databases (Medline, Global Health, Embase, Social Policy, and Web of Science) yielded 2167 abstracts, and resulted in the inclusion of 118 full-text articles with 1122 estimates of the prevalence of intimate partner homicide after double-blind screening. All studies were included that reported the number or proportion of women or men who were murdered by an intimate partner in a country, province, or town, using an inclusive definition of an intimate partner. Additionally, a survey of official sources of 169 countries provided a further 53 estimates. We selected one estimate per country-year using a quality assessment decision algorithm. The median prevalence of intimate partner homicide was calculated by country and region overall, and for women and men separately. FINDINGS: Data were obtained for 66 countries. Overall 13·5% (IQR 9·2-18·2) of homicides were committed by an intimate partner, and this proportion was six times higher for female homicides than for male homicides (38·6%, 30·8-45·3, vs 6·3%, 3·1-6·3). Median percentages for all (male and female) and female intimate partner homicide were highest in high-income countries (all, 14·9%, 9·2-18·2; female homicide, 41·2%, 30·8-44·5) and in southeast Asia (18·8%, 11·3-18·8; 58·8%, 58·8-58·8). Adjustments to account for unknown victim-offender relationships generally increased the prevalence, suggesting that results presented are conservative. INTERPRETATION: At least one in seven homicides globally and more than a third of female homicides are perpetrated by an intimate partner. Such violence commonly represents the culmination of a long history of abuse. Strategies to reduce homicide risk include increased investment in intimate partner violence prevention, risk assessments at different points of care, support for women experiencing intimate partner violence, and control of gun ownership for people with a history of violence. Improvements in country-level data collection and monitoring systems are also essential, because data availability and quality varied strongly across regions. FUNDING: WHO, Sigrid Rausing Trust, and the UK Economic and Social Research Council.