INTRODUCTION: The purpose of this study was to identify trajectories of older adults’ television viewing (TV) time over 12 years; and, to examine their associations with performance-based measures of physical function. METHODS: Data on TV time (hours/week) and socio-demographic factors were collected at each assessment of the Australian Diabetes, Obesity and Lifestyle (AusDiab) Study (1999/2000; 2004/2005; 2011/2012), with objective measures of physical function (2.44m timed-up-and-go [TUG, seconds] and knee extensor test [KES, kg]) collected at the final (2011/2012) assessment. Regression analyses examined predictors of trajectory membership and associations with TUG and KES in those aged 60+ years in 2011/2012. RESULTS: Six TV time trajectories were identified among the 1938 participants (aged 60-97, 54% female): consistently-low (9.7%); low-increasing (22.3%); moderate-decreasing (13.5%); moderate-increasing (30.3%); consistently-high (18.9%); and, high-increasing (5.2%). There were no statistically significant relationships with TUG (p>0.05). In the fully adjusted model, KES performance was significantly better in the consistently-low, low-increasing and consistently- high trajectories, compared to the moderate-increasing trajectory (P < 0.001, R = 0.33). CONCLUSION: 12-year trajectories of TV time were associated with muscle strength in older adults. These findings suggest that patterns of sedentary behavior can be a determinant of muscle strength in later life.