Objective: The World health organization (WHO) has accepted Keith Edward scoring system for the diagnosis of childhood tuberculosis (TB). In the present study, we evaluated this scoring system. Methods and Results: We included 53 children with confirmed TB involving different organs, admitted in NB Medical College, during two years period as cases; and 50 randomly selected, age, sex, and organ matched confirmed non-TB cases as controls. We noticed 15.1% false negative and 22% false positive results in our study, and the scoring system had 84.9% sensitivity, 78% specificity, and 80.36% positive predictive value. Likelihood ratio positive (LR+) was 3.86, likelihood ratio negative (LR-) was 0.19, and overall agreement was 81.55%. We observed that Keith Edward scoring system was less effective in children suffering from non-TB chronic diseases (false positive rate: 45.5%). We found no significant difference in nutritional status between study and control groups (P = 0.65). We noticed that more than 15-mm indurations for tuberculin test were specific for TB in children. Conclusion: We concluded that Keith Edward scoring system is good for public health purpose, but there is a scope for improvement, and further study is required for this purpose.