Falls in older adults are a substantial cause of injury, morbidity and mortality. Reducing the health burden of falls relies in part on identification of an individual’s risk of falling and targeted interventions that improve balance. In particular there is increasing evidence that many falls occur as a result of an impaired stepping response following loss of balance. It therefore follows that a thorough understanding of factors associated with impaired and successful step recovery is required to inform efforts to prevent falls via the design of more efficacious and safe exercise-based training interventions for fall prevention. The basic biomechanics (i.e. kinematics and joint moments) of balance recovery by stepping in older versus younger adults, and older fallers compared to non-fallers have been well described in the literature. The main focus of this thesis is to build on this fundamental understanding of balance recovery through the use of musculoskeletal models to understand the cause and effect relationship between muscles forces and the movement patterns and articular loads they generate. The purpose of the first experimental study (Chapter 3) was to determine the biomechanical factors that best predict maximal balance recovery performance. Forward loss of balance was achieved by releasing community-dwelling older adults (n = 117) from a maximal static forward lean angle from which participants attempted to recover balance by taking a single step. Lower limb strength measures, step recovery kinematics, stepping limb kinetics accounted for between 8 to 19%, 3 to 59%, and 3 to 61% of the variance in the Maximal Recoverable Lean Angle (MRLA) respectively.