The pupose of this study was to analyze influences on the bony union, to evaluate results and to consider effective methods of the treatment of infected nonunited fracture of the femur Seventeen patients who had infected nonunited fracture of the femur were managed from January 1989 to January 1996. We reviewed the results according to the method of treatment. Fixation were judged to be unstable in all of patients who had undergone primary internal fixation so that we treated them with radical debridement of soft tissue and necrotic bone. The bacterial cultures usually revealed a mixed infection and the organisms cultured from the infected fracture site were, in order of frequency , Staphylococcus aureus, Escherichia coli, Streptococcus, Pseudomonas, and Enterococcus. At the time of final evaluation, functional results according to criteria of Sanders-Swiontkowski-Helfet were excellent in three, good in five, fair in four, and poor in five. The average motion of the knee joint ranged from 6.7(0-20) of extension to 75.5(50-130) of flexion. The five patients had shortening of affected limb (mean 1.8±.2cm). Effective treatment of the infected non-union should be achieved not only bony union but also bacteriological and clinical remission of infection with subsequent closure of the wound and physical rehabilitation of the patient. Rigid internal fixation at the site of non-union can provide fracture healing as well as eradicate infection by improving the biologic environment.