Abstract Background Delayed tuberculosis (TB) diagnosis and treatment increase morbidity, mortality, expenditure, and transmission in the community. This study assessed patient and provider related delays to diagnosis and treatment of TB. Methods A cross-sectional study was conducted among 735 new adult TB cases registered between January to December 2015 in 10 woredas equivalent to districts of southwestern Ethiopia. Data were collected through face-to-face interview of patients within the first 2 months of treatment initiation. Delay in days was tracked at three intervals: between onset of symptoms and self-presentation (Patient delay), Self-presentation to treatment initiation (Provider delay) and total delay. Days elapsed beyond median were used to define the delays. Bivariate and multiple logistic regression models were fit to identify predictors of delays and statistical significance was judged at p