Purpose: To investigate longitudinal trajectories of vocational recovery (VR) among individuals with severe mental illness (SMI) who participated in the Danish Individual Placement and Support (IPS) trial, and whether the IPS intervention, socio-demographic and disease-specific characteristics predicted trajectory membership. Methods: In an observational study design, we used previously collected data from the Danish IPS trial (N = 720). VR was defined as ‘weeks in competitive employment or education in the past 6 months and was measured after 0.5, 1, 1.5, 2 and 2.5 years, using data from the Danish Register for Evaluation of Marginalization (DREAM) database. Latent growth mixture modelling in Mplus statistical software (version 7) was applied to identify trajectories of VR. Multinomial logistic regression analyses were used to estimate predictors for trajectory membership. Results: Four trajectories were identified: ‘Low VR’ (61.3%), ‘Low Increasing VR (8.2%), ‘Increasing Decreasing VR’ (7.2%) and ‘High VR’ (23.4%). Receiving the IPS intervention increased odds of membership in ‘High VR’ compared to ‘Low VR’ (OR = 2.18; 95% CI 1.37–3.48) and so did higher education (OR = 2.25; 95% CI 1.39–3.64), higher cognitive function (OR = 1.17; 95% CI 1.02–1.35), higher motivation to change (OR = 1.04; 95% CI 1.02–1.05) and previous work history (OR = 1.64; 95% CI 1.09–2.46). Higher age decreased odds of membership in the ‘High VR’ (OR = 0.95; 95% CI 0.93–0.98) compared to ‘Low VR’. Conclusion: There was high heterogeneity in the identified VR trajectories, despite that all participants expressed a desire for work and education at baseline. Improvements of the IPS intervention are needed to support specific groups in achieving and retaining employment.