Background: The proportion of births attended by skilled health personnel in Ethiopia is very low and the maternal mortality ratio currently is 676 per 100,000 live births. This study aimed at assessing level of institutional delivery service utilization and associated factors among mothers who gave birth during the last 12 months prior to the study in Hetosa district, Arsi zone Ethiopia. Methods: Community based cross sectional study design was conducted among 735 mothers who gave birth within last one year at Hetosa District in 2015. The collected data were entered into computer usingEpi-info version 3.5.1 and exported to SPSS version 20 software for analysis. Univariate, Bivariate and multivariate analysis were done. Significance level and association of variables was tested by using 95% confidence interval (C.I) and odd ratio and p-value less than 0.05 was taken as statistically significant. Result: Forty nine percent of the respondentsgave birth at health facilities and 98% delivered at public health facility. Out of those mothers who delivered at home, 36% were assisted by neighbor and the main reason for their home delivery was easy (precipitate) labor and very short duration which force mother to deliver at home 89% followed by transport problem (7.2%). According to multivariate regression analysis mothers who reside in urban, did not live with her husbandand had one live birthwere more likely to give birth at health institution when compared to their counter parts. Conclusion and recommendation: This study revealed that institutional delivery was high as compared to some of the studies conducted on different parts of the country. Policy makers and health care planners need to recognize the factors hampering institutional delivery and work on improving the situation.