We assessed the roles of energy intake and physical activity in the relationships among television (TV) viewing, body composition, and obesity using high-quality measurement methods. Adult women (n = 300) reported TV viewing behavior, which was categorized into infrequent (≤ 1 h/day), moderate (2 h/day), and frequent (≥ 3 h/day) viewing. Body fat percentage (BF%) was assessed using plethysmography (Bod Pod) and BMI was calculated from height and body weight. Energy intake and physical activity, including time spent in sedentary, moderate, and vigorous physical activity (PA), were objectively measured using 7-day weighed food records and 7-day accelerometry, respectively. The mean BF% of frequent TV viewers (34.6 ± 6.9%) was significantly greater (F = 3.9, P = 0.0218) than those of moderate (31.5 ± 6.7%) and infrequent viewers (30.8 ± 7.0%); however, BMI did not differ across the TV viewing groups (F = 0.8, P = 0.4172). Controlling statistically for differences in age, education, time in sedentary activity, time in moderate activity, and energy intake, considered individually, had no influence on the relationships between TV viewing and BF%, nor TV and BMI. Moreover, the relationship between TV and BF% remained significant after adjusting for differences in BMI (F = 3.6, P = 0.0276). However, adjusting for total PA reduced the relationship between TV and BF% to nonsignificance (F = 2.5, P = 0.0810), as did time spent in vigorous PA (F = 2.2, P = 0.1307). These data suggest a strong relationship between TV viewing and BF%. This association appears to be due, in part, to differences in total PA, particularly vigorous PA, but not time spent in sedentary activity, moderate activity, or energy intake.