Using panel data extracted from wave 1, 2, 4, 5 and 6 of the Survey of Health, Ageing and Retirement in Europe, this paper investigates the impact of unemployment on health outcomes of the European elderly during the period 2004 to 2015. Various estimators with panel fixed-effects are employed to control for the time-invariant unobserved confounders. To eliminate selection bias, the study separates the impacts of plant closures and being laid off from other potentially endogenous reasons for unemployment. Most importantly, finite mixture models are used to account for unobserved heterogeneity. Results from the models using a dummy for unemployment show significant impacts on self-rated health, chronic disease, and mental health. Models with plant closures and being laid off as two exogenous variables confirm the similar impacts on mental health. The impacts of plant closures and being laid off are slightly higher for chronic disease, but only being laid off worsens self-rated health. Notably, women are more likely to get chronic disease when they lose their jobs by plant closures, even once a year. Unemployment also damages women's mental health but the marginal effects for men are even larger. The finite-mixture models separately estimate the health impacts on two latent classes and show that members of the class with more ex-post ill-health tend to suffer more health problems due to unemployment. Age, marriage, number of children, education and industry sectors jointly determine the class membership.