Falls can have severe consequences for elderly adults. In 2000, nearly 10,300 people aged 65 years or older died as a result of falls, and 2.6 million individuals were treated for non-fatal fall-related injuries. In order to reduce fall incidences, it is important to identify possible causes of falls, such as muscle weakness and imbalance. In this study, we examined balance control in the elderly during task transitions while performing the Timed Up and Go test (TUG). The TUG is a commonly used clinical balance test that includes transition phases between three daily activity tasks: sit-to-stand, walking and turning. Our findings suggested that elderly adults, especially fallers, have reduced balance control ability while making transitions during TUG. During sit-to-walk (STW), when compared to young adults, elderly adults demonstrated a smaller forward center of mass (COM) velocity, a smaller anterior-posterior (A-P) COM-Ankle angle, and a larger upward kinetic energy ratio at seat-off. Additionally, the medial-lateral COM control in elderly fallers was also perturbed due to their significant reduction in forward COM velocity. The reduced initial hip extensor moment and increased ankle plantarflexor moment in elderly fallers was associated with their reduced generation of horizontal momentum during STW. Smaller A-P COM-Ankle angles and taking more steps when making a turn demonstrated a reduction in balance control ability in elderly adults. Our analyses suggest that balance control is an important factor contributing to longer STW and turning durations of TUG. Furthermore, lower extremity muscle strength at hip and knee joints demonstrated a stronger association with STW than turning duration. To enhance the early detection of fall risk, we also assessed the ability of balance tests to predict future risk of falling in elderly adults. Our results indicated that biomechanical balance parameters measured during TUG were associated with future fall status. Among all biomechanical parameters investigated, frontal plane balance control parameters appear to be the most significant predictors for future falls. This dissertation includes unpublished co-authored material.