Abstract Background It is not known how ethnic differences in COVID-19 deaths in the Netherlands evolved throughout the pandemic, especially after introduction of ethnicity-oriented COVID-19 prevention measures. We investigated associations between ethnicity and COVID-19 deaths across first wave of the pandemic, inter-wave period, and second wave in the Netherlands. Methods We obtained multiple registry data from Statistics Netherlands spanning from 01 March 2020 to 14 March 2021 comprising of 17.4 million inhabitants. We estimated incidence rate ratios (IRRs) for COVID-19 deaths among ethnic groups using Poisson regression models and adjusted for relevant socio-demographic factors. We used similar models to estimate IRRs for non-COVID-19 deaths among ethnic groups. Results Ethnic minority populations exhibited higher risk of COVID-19 deaths than the Dutch origin population throughout various study periods. The most elevated risk of COVID-19 deaths was in populations originating from low- and middle-income countries, especially those with Turkish, Moroccan, and Surinamese background. The elevated risk of COVID-19 deaths among ethnic minority groups (as compared to Dutch origin population) was higher in inter-wave period (4 times higher) and second wave (2 times higher) when compared to the first wave (1.5 times as higher). Ethnic differences in COVID-19 deaths were larger compared to non-COVID-19 deaths. Conclusions Ethnic differences in COVID-19 deaths persisted across first wave, inter-wave period and second wave in the Netherlands despite introduction of ethnicity-oriented prevention measures. Research on explanatory mechanisms and novel prevention measures are needed to address the ongoing differences in COVID-19 deaths across ethnic groups. Key messages • Ethnic differences in COVID-19 deaths persisted in the Netherlands despite introduction of ethnicity-oriented prevention measures.We therefore call for better prevention measures. • Well known drivers of SARS-CoV-2 infection such as household wealth, did ...