Retirement from work is a major life event requiring adaptation to new life circumstances. The resource-based dynamic model of retirement adjustment suggests that well-being will change due to changes in individual resources. In the present study, we test this hypothesis by investigating longitudinal and bidirectional associations between life satisfaction and perceived resources (i.e., self-esteem, autonomy, social support, self-rated physical health, self-rated cognitive ability, and financial satisfaction) over a 4-year period in the transition from work to retirement. Our sample included annual assessment data from 497 older adults (aged 60–66) in the population-based HEalth, Ageing, and Retirement Transitions in Sweden (HEARTS) study. Results from bivariate latent change score models showed weak but consistent associations between changes in perceived resources and changes in life satisfaction over the retirement transition. Analyses of cross-lagged effects also revealed bidirectional associations. Self-esteem, self-rated physical health, and total resource capability were positively related to changes in life satisfaction, and life satisfaction was positively related to changes in self-esteem, autonomy, self-rated physical health, and self-rated cognitive ability. The total resource capability accounted for 12% of the changes in life satisfaction in the first years following retirement. Life satisfaction accounted for 16% of the changes in autonomy in the transition from work to retirement. Our findings demonstrate that perceived resources are important for life satisfaction in the retirement transition, at the same time as overall life satisfaction accounts for how we perceive and evaluate our own resources during this process.