The Demand-Control-Support (DCS) and the Effort-Reward Imbalance (ERI) models assess different psychosocial factors. This study investigates whether a combination of these models increases their ability to predict sickness absence, as compared to results based on each model separately. A cross-sectional study with nursing personnel (N = 1307) in Brazil was performed. Regression analyses were conducted in three stages: analysis of each scale of the models and sickness absences; assessment of the independent association of each model with sickness absences; assessment of the associations of three combinations of models/scales with sickness absences: DC and social support (SS), ERI and overcommitment, and DC and ERI. As regards comparisons between the stress models, ERI was shown to be independently associated with short (up to 9 days) and long (10 days or more) spells of absenteeism. The same result held true for low social support. The combinations DC-ERI and DC-SS were better predictors for short spells than each model/scale separately, whereas for long spells, the combination DC-SS was the best predictor. ERI seems to be a good instrument for predicting absenteeism if used alone, whereas DC performed better when combined with ERI or SS. An improved risk estimation of sickness absences by combining information from the two models was observed.