We investigated the relationship of the retraction force to the location of force application, retarding force and bracket width during simulated sliding tooth movement along an arch wire. Point 1 for retraction was located at the center of the bracket, and points 2 and 3 were at 4.0 mm and 6.0 mm from the bracket slot, respectively. Weights of 100 gm, 200 gm, and 400 gm were suspended at 9.0 mm from the bracket slot as the point of simulated center of resistance. Stainless steel standard edgewise wide, medium, and narrow twin brackets were engaged with two elastomeric ligatures on a stainless steel wire (0.016 x 0.016 inch). The bracket was retracted at the rate of 0.1 mm per second for a distance of 2.0 mm. Measurements were repeated six times, and the results were compared with multiple ANOVA tests. For all brackets, with an increase of the retarding weight, the mean retraction force at points 1 and 2 increased but decreased at point 3. The mean retraction force at point 1 for the narrow twin bracket was significantly (p < 0.05) higher than that for the wide twin bracket at all retarding force levels. However, the mean retraction force at points 2 and 3 for the narrow twin bracket was significantly (p < 0.05) lower than for the wide twin bracket at all retarding force levels. These findings indicated that the point of force application, the resistance force of a tooth, and the width of the bracket are crucial in consideration of the tipping moments on the bracket.