Background As the prevalence of chronic disease amongst older workers is high and increasing, it is important to know if the large subgroup of older workers with chronic disease has specific needs when it comes to prolonging participation in paid work. Objectives To investigate differences and similarities in predictors of having paid work in workers aged 55? with and without chronic disease. Methods Workers aged 55-62 years were selected from the 2002-2003 cohort of the Longitudinal Aging Study Amsterdam (n = 333). Potential predictors were: health, personality, work characteristics, and demographics. Per potential predictor, a logistic regression coefficient for 'having paid work in 2005-2006' was calculated for workers with and without chronic disease. A pooled estimate was computed and differences between the pooled estimate and the coefficients were tested. Results Follow-up data were available for 95 %, of whom 67 % still had paid work. Predictors of having paid work were similar for workers with and without chronic diseases, except for physical workload (v 2 = 5.37; DF = 1) and psychosocial resources at work (v 2 = 5.94; DF = 1). Having more psychosocial resources (OR = 3.57; 95 %CI 1.33-10.0) was predictive for having paid work in workers with chronic disease and not in workers without chronic disease. Lower age, more weekly working hours, no functional limitations, fewer depressive symptoms, lower neuroticism scores, and more sense of mastery were significantly associated with having paid work in all workers. Conclusions Differences between predictors of having paid work between workers with and without chronic disease should be taken into account when aiming to prevent exit from the workforce. In particular the vulnerable subgroup of older workers with chronic disease and low psychosocial resources at work is more likely to quit working.