**Background:** Capturing the complexity of family life courses as predictors of later-life outcomes like wealth is challenging. Previous research has either (a) assessed a few selective but potentially irrelevant summary indicators, or (b) examined entire life course clusters without identifying specific important aspects within and between them.
**Objective:** To investigate which family life-course variables—encompassing variables that capture the order, duration, and timing of states and transitions—are key personal wealth predictors for Western Germans aged 50 to 59. And analyse the strength and direction of associations between relevant variables and personal wealth, and whether these differ by gender.
**Methods:** We used German Socio-economic panel study (SOEP) data and combined feature selection, sequence analysis tools, and regression techniques.
**Results:** We identified 23 family life-course variables as relevant predictors, with two—the time spent never-married, both without and with children—deemed most relevant. Most family life-course variables were negatively associated with personal wealth and characterised by single parenthood, marital separation or early marital transitions with or without fertility transitions. The prevalence and significance of associations between these variables and personal wealth differed partly across genders. Results highlight the importance of previously concealed family life-course variables for wealth inequalities in late working age.
**Contribution:** We extend previous research on the nexus between family demography and wealth stratification by using a novel, data-driven approach that more effectively explores family life-course complexities by considering the ‘entire’ universe of variables that describe such life courses and identifying those life-course variables that are relevant wealth predictors.
**Keywords:** Family dynamics, Life course, Social stratification, Wealth, Feature Selection
