Background The aim of this study was to assess the predictive validity of the Patient and Observer Assessment Scale (POSAS), in order to determine whether it can be used to predict final scar quality. Methods Patients with a maximum TBSA burned of 20% who were treated in a Dutch burn center and participated in two scar assessments at 3 months and >18 months post-burn were included. Scar quality assessment consisted of the POSAS, Dermaspectrometer® (color) and Cutometer® (elasticity). Predictive validity was determined in three ways: (1) the discriminative ability to distinguish good from reduced long term scar quality, (2) correlations between POSAS items score at the two subsequent assessments and (3) linear regression was conducted to identify POSAS items as independent predictors. Additionally, reliability, construct validity and interpretability were assessed. Results A total of 141 patients were included with a mean TBSA burned of 5.2% (±4.5). The ability of the Patient scale to discriminate between good and reduced long term scar quality was adequate with an area under the curve (AUC) of 0.728 (CI 0.640–0.804), the ability of the Observer scale was good with an AUC of 0.854 (CI 0.781–0.911). Correlations between items scored T3 and T > 18 were at least adequate. On item level, pain and stiffness (Patient) and pliability and relief (Observer) were identified as significant predictors for reduced long term scar quality. The POSAS was reliable, construct validity was adequate at three months but declined at?>18 months. Conclusion This study found that final scar quality can be adequately predicted by an early POSAS assessment at three months.