Background Cognitive impairment is a prevalent, independent marker of readmission in heart failure (HF), but the screening is time-consuming. This study sought (1) to identify HF patients at low risk of cognitive impairment (obviating screening) and (2) to simplify a predictive model of HF outcomes by only using cognitive domains that are most predictive. Methods and Results The Montreal Cognitive Assessment was performed in 1152 Australian patients with HF who were followed for 12 months. One-third (376/1152) of the patients were enrolled into an HF disease management plan to reduce early readmission. Postdischarge outcomes in HF included 30- and 90-day readmission or death and days alive and out of hospital within 12 months of discharge. Cognitive impairment-present in 54% of patients-independently predicted HF outcomes. Normal cognition could be predicted with common clinical and sociodemographic factors with good discrimination (C statistic=0.74 [0.69-0.78]). The visuospatial/executive and orientation domains were most predictive of HF postdischarge outcomes. Using either Montreal Cognitive Assessment score or these 2 domains provided similar incremental values ( P=0.0004 and P=0.0008, respectively) in predicting HF outcomes (both C statistic=0.76) and could similarly identify a group of high-risk patients who benefited most from an HF disease management plan. Conclusions Cognitive function independently predicts HF outcomes and may also contribute to how a patient responds to intervention. The time and resources spent on cognitive assessment for risk-stratification in HF may be minimized by (1) identifying patients with low risk of cognitive impairment and (2) simplifying the screening instrument to include only the domains that are most predictive of postdischarge outcomes in HF.