We examined the impact of COVID-19 pandemic onset (2020 April) on homicide mortality in the United States. We conducted a single interrupted time series analysis using homicide events from the National Vital Statistics System that occurred over six years (2017-2022), with COVID-19 onset as an interruption. Monthly homicide deaths rates were calculated per 100,000 person-years to create a monthly time series. We used autoregressive integrated moving average regression, adjusted for seasonality, to model the immediate and sustained trend changes in the homicide mortality rate ratios due to the pandemic. We stratified models by length of stay-at-home order, race and ethnicity, sex, age, and weapon used to examine effect measure modification. In Jan 2017, the US homicide mortality rate was 5.9/100,000 PY. While there were annual seasonal changes, the overall time trend before April 2020 was stable. However, with COVID-19 onset, the overall homicide mortality rate ratio increased by 32% (95% CI: 0.23, 0.41), which persisted through 2022 without additional trend changes, but with seasonal variations. Immediate increases with stable sustained trends in homicide rates were also observed in most stratified analyses. COVID-19 pandemic onset is associated with US homicide mortality rates immediately increasing and remaining stable and higher afterwards.