The purpose was to determine the relationship between insulin resistance (IR) and risk of gaining body fat percentage (BF%), body weight, and abdominal fat over 18 months. A prospective cohort study was conducted using a sample of 226 women. IR was assessed using fasting blood insulin and glucose levels to calculate homeostatic model assessment (HOMA). Participants were divided into High (4th quartile) Moderate (2nd and 3rd quartiles), and Low (1st quartile) HOMA categories. BF% was estimated using plethysmography (Bod Pod), weight was measured in a standard swimsuit, and abdominal fat was indexed using the average of two circumferences taken at the umbilicus. Participants wore accelerometers and completed weighed food logs for 7 consecutive days to control for the effect of physical activity (PA) and energy intake, respectively. On average, women in the High HOMA group decreased in BF% (-0.48 ± 3.60), whereas those in the Moderate (0.40 ± 3.66) and Low HOMA (1.17 ± 3.15) groups gained BF% (F = 5.4, P = 0.0211). Changes in body weight showed a similar dose-response relationship (F = 4.7, P = 0.0317). However, baseline IR was not predictive of changes in abdominal fat (F = 0.8, P = 0.3635). Controlling for several covariates had little effect on gains in BF% and weight, but adjusting for initial BF% and/or initial weight nullified changes in BF% and weight across the IR groups. In conclusion, women with High HOMA tend to gain significantly less BF% and weight than women with low or moderate HOMA. The decreased risk appears unrelated to several covariates, except initial BF% and weight levels, which seem to play key roles in the relationships.