Abstract: Objective To establish a prediction model of short-term recurrence of cured initial-smear-positive pulmonary tuberculosis.  Methods This research was a prospective longitudinal study. 3293 cured initial-smear-positive pulmonary tuberculosis cases from Jan.2005 to Dec. 2006 from Guangzhou City were followed-up for 2 years. Influencing factors of recurrence of cured initial-smear-positive pulmonary tuberculosis were explored from clinical and social aspects and a logistic regression model for prediction were developed.  Results The bacteriology recurrence rate is 4.6%. Short-term recurrence prediction model is logit P=-9.546+0.586X₁+1.959X₆+1.449X¹⁰+1.118X¹⁵+1.027X¹⁶+1.525X¹⁸+0.674X²²+2.042X²³ with good prediction Results .  Conclusion s The prediction model provided a scientific evidence-based basis to reduce the recurrence of tuberculosis and a targeted instruction for the control of tuberculosis epidemic.

Key words: tuberculosis, pulmonary, recurrence, forecasting