Introduction Prison settings have been often identified as important but neglected reservoirs for TB. This study was designed to determine the prevalence of undiagnosed pulmonary TB and assess the potential risk factors for such TB cases in prisons of the Tigray region. Method A cross-sectional study was conducted between August 2013 and February 2014 in nine prisons. A standardized symptom-based questionnaire was initially used to identify presumptive TB cases. From each, three consecutive sputum samples were collected for acid-fast bacilli (AFB) microscopy and culture. Blood samples were collected from consented participants for HIV testing. Result Out of 809 presumptive TB cases with culture result, 4.0% (95% CI: 2.65-5.35) were confirmed to have undiagnosed TB. The overall estimated point prevalence of undiagnosed TB was found to be 505/100,000 prisoners (95% CI: 360-640). Together with the 27 patients who were already on treatment, the overall estimated point prevalence of TB would be 793/100,000 prisoners (95% CI: 610-970), about four times higher than in the general population. The ratio of active to passive case detection was 1.18: 1. The prevalence of HIV was 4.4% (36/809) among presumptive TB cases and 6.3% (2/32) among undiagnosed TB cases. In a multivariate logistic regression analysis, chewing Khat (adjusted OR = 2.81; 95% CI: 1.02-7.75) and having had a close contact with a TB patient (adjusted OR = 2.18; 95% CI: 1.05-4.51) were found to be predictors of undiagnosed TB among presumptive TB cases. Conclusions This study revealed that at least half of symptomatic pulmonary TB cases in Northern Ethiopian prisons remain undiagnosed and hence untreated. The prevalence of undiagnosed TB in the study prisons was more than two folds higher than in the general population of Tigray. This may indicate the need for more investment and commitment to improving TB case detection in the study prisons.