Objective To verify a method of calculation and prediction by screening the antibacterial components of classical Chinese medicinal herbs (TCM) for anti tuberculosis, and preliminarily evaluate the anti tuberculosis effect of the selected components. Methods The components database of Stemona, Bletilla striata and Chuanbei was established. Dock method was used to predict the anti tuberculosis components from the database, and then mycobacterium smegmatis, MIC and inhibition zone methods were used to evaluate the anti-tuberculosis effect of these predicted components. Results Three ingredients were selected. In vitro experiments also showed that the selected ingredients had the effect of anti-mycobacterium smegmatis by MIC and inhibition zone . Conclusions The virtual screening method could decrease consumption and increase the efficiency of finding anti-tuberculosis ingredients of TCM.