Abstract Body: There are various pre-clinical models of heart failure with preserved ejection fraction (HFpEF) designed to model different clinical phenotypes of HFpEF. However, detailed assessment and head to head comparison of diastolic function by invasive and non-invasive techniques in these models is lacking. Our goal was to comprehensively evaluate cardiac structure and function in three HFpEF mouse models – deoxycortisterone acetate (DOCA)-salt, Nω-nitro-l-arginine methyl ester high fat diet (LNAME+HFD), and aged female mice fed HFD with angiotensin infusion (Ang+HFD) – using both non-invasive and invasive methods. We employed pulse wave and tissue Doppler echocardiography, cardiac magnetic resonance imaging (CMR), and left ventricular (LV) catheterization in the commonly used C57BL/6J mouse, enabling direct comparisons between models. Cardiac hypertrophy was assessed by heart weight (HW) to tibia length (TL) ratio. The Mann-Whitney test was used for statistical testing within each model. We found that on echocardiography, all models demonstrated diastolic dysfunction with significant differences (p < 0.05) or trends for significance (p < 0.1) on tissue doppler peak mitral annular velocity during early filling (e’), ratio of peak early mitral flow velocity (E) to e’ (E/e’), ratio of peak early mitral flow velocity (E) to e’ (E/e’), isovolumetric relaxation time (IVRT), and E pressure half time (E PHT). CMR was notable for findings of reduced right ventricular ejection function in the LNAME+HFD group (36.1±2.7% vs 49.8±1.4% in chow controls, p<0.01). On invasive hemodynamics, only DOCA-salt had a higher LV diastolic time constant (t, 11.7±1.4 vs 6.9±0.8 ms in sham controls, p=0.02) and only Ang+HFD had a higher minimum rate of pressure change in the ventricle (dP/dt_min, -3506±342 vs -5268±227 in chow fed, vehicle infusion controls, p<0.01). No model showed a difference in LV end diastolic pressure. For cardiac hypertrophy, we found that both DOCA-salt (8.1±0.2 vs 10.3±0.5, p<0.01) and Ang+HFD (9.2±0.3 vs 11.5±0.4, p<0.01) models demonstrated increased HW to TL ratio, while LNAME+HFD did not. These results highlight differences in cardiac structure and function in HFpEF mouse models and the need for multi-modal assessment in pre-clinical studies.