This study investigates whether increasing levels of job autonomy are linearly associated with improved psychological well-being, or whether the association is non-linear with higher levels of job autonomy being negatively associated with psychological well-being. The study is based on a survey in 4,340 individuals. After six months a follow-up survey was conducted. Associations were analysed using spline models in cross-sectional and prospective analyses of the full study population. We stratified the study population in four types of work for further cross-sectional analyses. The cross-sectional analysis of the full study population indicated a non-linear relationship as the association attenuated but remained positive at all levels of job autonomy. The prospective analysis indicated a positive linear association. The stratified cross-sectional analyses indicated a linear association in three of the four types of work and a non-linear but positive association among respondents in client-related work. Findings are supported in sensitivity analyses using alternative outcomes: vigour, job satisfaction and work stress. We conclude that job autonomy is positively associated with psychological well-being. Findings have relevance for job redesign by indicating that higher levels of job autonomy are beneficial for the psychological well-being of workers – especially among workers with low levels of job autonomy.