The number of older people who experience marital break-up has increased in many Western countries. However, limited empirical attention has been given to the study of the consequences of later-life divorce or separation. Previous studies on gray divorce are often cross-sectional and tend to capture a mix of short- and long-term effects of divorce and possibly selection effects into divorce. Drawing on data from nine waves of the UK Household Longitudinal Study (2009/2010-2017/2018), we analyze the effect of marital break-up on the mental health of 909 adults aged 50 or over to test the crisis model and the chronic strain model of divorce. We use fixed effects linear regression models to account for time-invariant confounders and distinguish between pre- and post-divorce effects. Our results indicate that older adults' depressive symptoms (GHQ) increase in the years before and upon union dissolution. After separation, depressive symptoms decrease and return to approximately previous baseline levels. Our analyses on heterogeneity in the effects of gray divorce show that post-divorce adjustment is faster for childless adults than for parents. We find no evidence that adjustment after gray divorce is slower for women than for men, or for persons who already experienced a prior union dissolution than for those who separate for the first time. The results are consistent with the crisis model of divorce but in contrast with the chronic strain model of divorce. Older adults are able to adjust to marital break-up, and their fertility histories tend to moderate the negative effect of later-life divorce on mental health.