All types of aerobic exercise are assumed to affect cardiovascular risk similarly. There are few studies of swimming, but complex responses to water-based exercise suggest its potential for differential effects. The aim of the study was to compare the effects of swimming and walking on fitness, body weight, lipids, glucose, and insulin in older women. Sedentary women aged 50 to 70 years (N = 116), randomly assigned to swimming or walking plus usual care or a behavioral intervention, completed 3 sessions per week of moderate-intensity exercise, supervised for 6 months then unsupervised for 6 months. After 6 months, 1.6-km walk time decreased in walkers and swimmers, with greater improvement in walkers (1.0 vs 0.6 minute, P = .001). In swimmers, but not walkers, distance swum in 12 minutes increased (78.1 vs -2.2 m, P = .021). Waist and hip circumferences (80.8 vs 83.1 cm and 101.8 vs 102.4 cm; P = .023 and P = .042, respectively) and insulin area under the curve (oral glucose tolerance test) (5128 vs 5623 μU/[L 120 min], P < .05) were lower with swimming. Lipids did not differ between groups. At 12 months, fitness was maintained. Relative to walking, swimming reduced body weight by (1.1 kg, P = .039) and resulted in lower total and low-density lipoprotein cholesterol (0.3 and 0.2 mmol/L; P = .040 and P = .049, respectively). The magnitude of the difference in the reduction of insulin area under the curve between swimming and walking was greater at 12 months; however, the significance was attenuated (4677 vs 5240 μU/[L 120 min], P = .052). Compared with walking, swimming improved body weight, body fat distribution, and insulin in the short term and, in the longer term, body weight and lipid measures. These findings suggest that the type of exercise can influence health benefits.