Heart failure and atrial fibrillation (AF) often coexist and are associated with diastolic dysfunction (DD). DD involves abnormalities in mechanical function during the diastolic phase and progresses over time. AF, the most common cardiac arrhythmia globally, increases the risk of ischemic events. Zones with reduced velocity or blood stasis are classified as pro-thrombotic regions, also correlated with endothelial signaling for cardiovascular remodeling. Computational Fluid Dynamics (CFD) allows to evaluate complex hemodynamic patterns associated with the anatomical intricacies of the left atrium (LA). In this context, the main goal of the paper was to use CFD to assess LA hemodynamics focusing in blood stasis. Different clinical scenarios, including a healthy condition and the three stages of DD (altered relaxation, pseudonormal, restrictive pattern), along AF, were numerically evaluated. The results showed distinct intra-atrial hemodynamics. In the healthy case, atrial contraction directed blood from the LA appendage (LAA) apex towards its ostium, reducing thrombus risks. In DD cases, stagnation zones were identified, worsening with DD severity. Altered relaxation preserving systolic peak also showed predominant flow from LAA apex to ostium, contrasting with increased blood stasis zones in DD II and DD III, attributed to reduced systolic peak and poor blood washout inside the LAA, increasing thrombogenesis risks. Thrombus-prone areas, typically assessed using Time-Averaged Wall Shear Stress indicators are difficult to obtain clinically, especially in AF patients. Regions characterized by low velocity (stasis zones) were identified to become larger with DD progression, potentially serving as additional markers of thrombotic propensity, as velocity data can be promptly obtained by ultrasound imaging in clinical practice.