Body composition is an important component of an individual's health and fitness profile. The purpose of this study was to assess the accuracy of bioelectrical impedance analysis (BIA) in detecting percent body fat (%BF) changes throughout a physical activity intervention in previously sedentary, overweight men and women. Fourteen men (mean ± SD; 46 ± 7 yr) and 22 women (47 ± 5 yr) with a body mass index of ≥25 kg·m−2 participated in an intervention of 10,000 steps per day for 36 wk. %BF was measured by air displacement plethysmography (ADP; criterion measure) and BIA at baseline, 20 wk, and 36 wk. During the 36-wk intervention, the women lost 1.6 ± 4.8 kg and the men lost 3.9 ± 4.2 kg. There was a significant three-way interaction for %BF across time points, device, and sex (P = 0.019). Among women, ADP and BIA detected %BF changes of −1.9% ± 2.0% and 0.5% ± 1.8%, respectively, from baseline to 20 wk, and −1.5% ± 2.3% and 0.3% ± 2.5%, respectively, from baseline to 36 wk. Among men, ADP and BIA detected %BF changes of −1.9% ± 1.8% and −1.5% ± 1.9%, respectively, from baseline to 20 wk, and −2.8% ± 3.7% and −2.3% ± 2.6%, respectively, from baseline to 36 wk. There was a significant difference in the accuracy of BIA in detecting %BF changes compared with ADP in women but not men on a group level. However, on an individual level, there was greater variation in BIA's accuracy in tracking %BF changes in both sexes. Compared with ADP, BIA is a suitable method for detecting %BF changes in a group of overweight men, but not women. However, caution should be exercised when using BIA to track body composition changes in either sex on an individual level.