Based on biologically plausible mechanisms and previous research, it is possible to hypothesize a reciprocal association between sustained attention and loneliness. We investigated this association using a cross-lagged modeling approach. Using data from 6,239 participants aged over 50 in TILDA, a nationally representative study of aging, we used structural equation models to investigate potential cross-lagged associations between sustained attention and loneliness, measured at baseline and again after four years. Sustained attention at baseline had a small association with loneliness four years later, but loneliness at baseline was not associated with sustained attention at follow-up. Auto-regressive associations were strong for both loneliness over time and sustained attention over time. Sustained attention may account for a small proportion of the variance in loneliness over time among older adults, and may constitute a risk factor in the development of loneliness. Implications for the identification of at-risk individuals and the prevention of loneliness are discussed.