Tuberculosis (TB) is one of the major public health and socio-economic issues in the 21st century globally. Assessment of TB treatment outcomes, and monitoring and evaluation of its risk factors in Directly Observed Treatment Short Course (DOTS) are among the major indicators of the performance of a national TB control program. Hence, this institution-based retrospective study was conducted to determine the treatment outcome of TB patients and investigate factors associated with unsuccessful outcome at Dilla University Referral Hospital, southern Ethiopia. Five years (2008 to 2013) TB record of TB clinic of the hospital was reviewed. A total 1537 registered TB patients with complete information were included. Of these, 942 (61.3%) were male, 1015 (66%) were from rural areas, 544 (35.4%) were smear positive pulmonary TB (PTB+), 816 (53.1%) were smear negative pulmonary TB (PTB-) and 177(11.5%) were extra pulmonary TB (EPTB) patients. Records of the 1537 TB patients showed that 181 (11.8%) were cured, 1129(73.5%) completed treatment, 171 (11.1%) defaulted, 52 (3.4%) died and 4 (0.3%) had treatment failure. The overall mean treatment success rate of the TB patients was 85.2%. The treatment success rate of the TB patients increased from 80.5% in September 2008-August 2009 to 84.8% in September 2012-May 2013. Tuberculosis type, age, residence and year of treatment were significantly associated with unsuccessful treatment outcome. The risk of unsuccessful outcome was significantly higher among TB patients from rural areas (AOR = 1.63, 95% CI: 1.21-2.20) compared to their urban counterparts. Unsuccessful treatment outcome was also observed in PTB- patients (AOR = 1.77, 95% CI: 1.26-2.50) and EPTB (AOR = 2.07, 95% CI: 1.28-3.37) compared to the PTB+ patients. In conclusion, it appears that DOTS have improved treatment success in the hospital during five years. Regular follow-up of patients with poor treatment outcome and provision of health information on TB treatment to patients from rural area is recommended.