A two-dimensional, sagittally-symmetric biomechanical model was developed to analyze the joint moments required to stabilize the trunk in a seated, dynamic, weight-moving task. Kinematic and reaction force data were measured while subjects moved a hand-held weight (0-4 kgf) at shoulder level to and fro at 1 Hz. These data were then used for model input and validation purposes. A second, simpler model was used to simulate how joint loads varied with weight held, trunk inclination, and movement frequency. The results for this seated task demonstrate a) significant trunk, hip, knee, and ankle joint moments (37, 13, 4, 13 percent of maximum strength values, respectively) were required, b) considerable intersubject differences in mean joint moments (more than 66 percent) were found, which primarily were due to subtle differences in body segment kinematics and lower extremities use, and c) the important role of the lower extremities in stabilizing the trunk in the seated posture.