The non-point loss of a pesticide through runoff and leaching from agricultural fields depends not only on its chemical properties but also on environmental factors such as soil type, rainfall intensity, and water table level. Two major agricultural calcareous soils in South Florida, Krome (loamy-skeletal, carbonatic, hyperthermic Lithic Udorthents) very gravelly loam (Krome) and Biscayne (loamy, carbonatic, hyperthermic, shallow Typic Fluvaquents) marl (marl), were packed in tilted stainless-steel runoff–leaching chambers for measuring the transport of endosulfan (6,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzadioxathiepin 3-oxide) isomers in leachates and surface runoff. Endosulfan loss was the greatest in runoff sediment, followed by loss in runoff water, and finally in the leachate. At a rainfall intensity of 75 mm h−1, the infiltration rate in the Krome soil was higher than in the marl soil. Endosulfan movement with infiltration water in Krome soil is a nonequilibrium process. Increasing rainfall intensity to 150 mm h−1 significantly decreased endosulfan leaching and increased runoff. High rainfall intensity and/or water table dramatically increased the endosulfan loss through the sediments in runoff. Marl soil has a higher endosulfan runoff potential than Krome soil. Endosulfan α had a lower average experimental distribution coefficient, KD, than endosulfan β (265 vs. 472). Loss of endosulfan α dissolved in runoff water and through leaching was higher than that of endosulfan β, while loss of endosulfan α from runoff sediments was lower than that of endosulfan β. Endosulfan concentrations in runoff water decreased exponentially with cumulative rainfall and runoff sediment.