Tuberculosis (TB) is an infection with global impact that over time demonstrates enormously high mortality rates. The vital need for improving novel and efficient anti-TB drugs is caused by the rising rate of appearance of Multi Drug Resistant (MDR) strains to the frequently utilized drugs.In addition, the longer periods of therapy and healing, mainly in the immune compromised patients aggrevates the situation. Recent studies indicate that computer-based techniques have been used successfully in the antibacterial research. In our current approach, utilizing combined pattern of computer-based methods as fragment-based de novo design, structure-based docking and scoring, in addition to similaritybased compound searching, led to introduce seven in silico designed compounds with probable antimycobacterial properties. Then, we investigated their potency against sensitive and resistant strains of Mycobacterium sp. in vitro. Findings resulted from antimycobacterial tests and MTT assay indicated that two compounds, 1-amino-4-(phenylamino) anthracene-9,10-dione and 5-fluoroindoline-2,3-dione have useful profile and maybe good candidates for developing novel antimycobacterial drugs.