Long‐distance commuting between home and the place of work is a means to combine both labour market participation and staying embedded in family and local social networks. But commuters often report high stress levels due to their commute and negative outcomes on their family life. Recent findings suggest that long commutes are harmful for partnership stability, especially if women commute. This contribution analyses the questions whether these findings can be replicated with a representative cohort panel study from Germany (pairfam), whether the expected negative outcome of female commuting is robust against partnership quality and labour division, and whether the risk of separation varies systematically across urban, suburban, and rural areas. Discrete event history models are applied to estimate the risk of separation for about 2,500 couples over 3 years (N = 49,408 person‐months). The findings indicate that women's long commuting is normally harmful for partnership stability but with one important exception: not if the couple lives in a suburban area. The article discusses whether the spatial variation in the influence of female commuting might be due to self‐selection processes in the course of family formation.