The Timed Up and Go (TUG) test is a widely used measure of mobility and fall risk among older adults that is typically scored using a stopwatch. We tested the hypothesis that a body-fixed accelerometer can enhance the ability of the TUG to identify community-living older adults with a relatively high fall risk of unknown origin. Twenty-three community-living elderly fallers (76.0 ± 3.9 years) and 18 healthy controls (68.3 ± 9.1 years) performed the TUG while wearing a 3D-accelerometer on the lower back. Acceleration-derived parameters included Sit-to-Stand and Stand-to-Sit times, amplitude range (Range), and slopes (Jerk). Average step duration, number of steps, average step length, gait speed, acceleration-median, and standard-deviation were also calculated. While the stopwatch-based TUG duration was not significantly different between the groups, acceleration-derived TUG duration was significantly higher (p = 0.007) among the fallers. Fallers generally exhibited lower Range and Jerk (p < 0.01). While TUG stopwatch duration successfully identified 63% of the subjects, an accelerometer-derived three-measure-combination correctly classified 87% of the subjects. Accelerometer-derived measures were generally not correlated with TUG duration. These findings demonstrate that fallers have difficulty with specific TUG aspects that can be quantified using an accelerometer. Without compromising simplicity of testing, an accelerometer can apparently be combined with TUG duration to provide complementary, objective measures that allow for a more complete, sensitive TUG-based fall risk assessment.