Background: Pregnancy loss is the terminated pregnancy before the completed pregnancy time. It is common in South Asia and sub-Saharan Africa. The main objective of thisstudy wasto identify the determinants and spatial distribution of pregnancy loss among zones of Ethiopia. Method: This study used the 2016 Ethiopia Demographic and Health Survey (EDHS), which was a community-based cross-sectional study design. All women of reproductive age within 5 years before the survey and having pregnancy history were the study population. The total women were 7,194. The spatial analysis and the multilevel models were performed using SAS 9.4 and STATA 17. Model performance was done with AIC and BIC and the smallest isrevealed the best model. Results: The prevalence of pregnancy lossin thisstud ywas 11.80%. The estimated intra-class correlation was 0.0765. Thisindicates about 7.65% of the total variation for pregnancy loss was due to the difference between zones whereas the remaining 92.35% of the variability to be accounted for individual and other unknown factors. The Ethiopian administrative zone Somali Zone 4, western Tigray, North Gondar, Mekele, southern Tigray, Eastern Tigray, Somali Zone 2, Wag Hemira, North Wollo, South Wollo, Oromia Zone, Somali unknown zone, Dega Habour, Hundene, Dire Dawa and western Gojjam were the higherrisk area of pregnancy loss. Conclusion: Thisstudy demonstrated that pregnancy losses vary across administrative zones in Ethiopia and both the hot-and cold spot areas were identified. Based on AIC and BIC binary multilevel model was preferred. The community husband education at the community level and residence, ANC, maternal work status, household number, wanted pregnancy, birth order, contraceptive, place delivery, Media Exposure, and BMI were significant factorsfor pregnancy loss. Keywords: pregnancy loss, multilevel, EDHS, ICC,spatial Analysis, Hotspot, Ethiopia