Objectives Diagnosing chronic heart failure (CHF) in general practice is challenging. Our aim was to investigate how general practitioners (GPs) diagnose CHF in real-world patients. Design Think-aloud study. Methods Fourteen GPs were asked to reason about four real-world CHF cases from their own practices. The cases were selected through a clinical audit. This was followed by an interview to get a deeper insight in their reasoning. The Qualitative Analysis Guide of Leuven was used as a guide in data analysis. Results We developed a conceptual diagnostic model based on three important reasoning steps. First, GPs assessed the likelihood of CHF based on the presence or absence of HF signs and symptoms. However, this approach had serious limitations since GPs experienced many barriers in their clinical assessment, especially in comorbid elderly. Second, if CHF was considered based on step 1, the main influencing factor to take further diagnostic steps was the GPs’ perception of the added value of a validated CHF diagnosis in that specific case. Third, the choice and implications of these further diagnostic steps (N-terminal pro B-type natriuretic peptide, ECG and/or cardiac ultrasound) were influenced by the GPs’ knowledge about these tests and the quality of the cardiologists’ reports. Conclusion This think-aloud study identified the factors that influenced the diagnostic reasoning about CHF in general practice. As a consequence, targets to improve this diagnostic reasoning were withheld: a paradigm shift towards an earlier and more comprehensive risk assessment with, among others, access to natriuretic peptide testing and convincing GPs of the added value of a validated HF diagnosis.