The purpose of this study was to identify the effects of mechanical traction applied to the knee joint on pain, knee range of motion (ROM), timed up and go (TUG) and Western Ontario and MacMaster Universities Osteoarthritis (WOMAC) in patients with knee osteoarthritis (KOA) of Kellgren-Lawrence radiological rating scale Ⅱ or Ⅲ. Twenty three patients participated in the experiment for a period of four weeks. After baseline assessment, the patients with KOA were randomized into two groups: the traction group (n1=12), which received traction with general physical therapy; and the control group (n2=11), which received general physical therapy only on unilateral knee joints. Patients received interventions once a day, three times a week, for four weeks. Wilcoxon signed-rank test was used to analyze the change of dependent variances within the group during pre and post intervention. Mann-Whitney U test was used to analyze the change of dependent variances as TUG and passive ROM between the two groups. Analysis of covariance was used to analyze the change of dependent variances as numeric pain rating scale (NPRS) and WOMAC score between the two groups. In Wilcoxon signed-rank test, the traction group improved significantly with regard to NPRS (p<.01), passive knee flexion ROM (p<.01), passive knee extension (p<.05), TUG (p<.01) and WOMAC scores (p<.01) after intervention for four weeks, but not for the control group. In the Mann-Whitney U test and analysis of covariance, no significant difference was seen among all the dependent variances after intervention for four weeks between the two groups. These outcomes suggest that further studies should be carried out to determine the effects of mechanical traction prior to using it for the treatment of patients with knee osteoarthritis.