Abstract Background Mothers’ self-efficacy for limiting their children’s television viewing is an important correlate of this behaviour in young children. However, no studies have examined how maternal self-efficacy changes over time, which is potentially important during periods of rapid child development. This study examined tracking of maternal self-efficacy for limiting young children’s television viewing over 15-months and associations with children’s television viewing time. Methods In 2008 and 2010, mothers (n = 404) from the Melbourne InFANT Program self-reported their self-efficacy for limiting their child’s television viewing at 4- and 19-months of age. Tertiles of self-efficacy were created at each time and categorised into: persistently high, persistently low, increasing or decreasing self-efficacy. Weighted kappa and multinomial logistic regression examined tracking and demographic and behavioural predictors of change in self-efficacy. A linear regression model examined associations between tracking categories and children’s television viewing time. Results Tracking of maternal self-efficacy for limiting children’s television viewing was low (kappa = 0.23, p