Although murine models for studying the development of cardiac dysfunction in diabetes mellitus are well established, their reported cardiac phenotypes vary. These reported divergences may, in addition to the severity of different models, also be linked to the methods used for cardiac functional assessment. In the present study, we examined the functional changes using conventional transthoracic echocardiography (in vivo) and isolated heart perfusion techniques (ex vivo), in hearts from two mouse models; one with an overt type 2 diabetes (the db/db mouse) and one with a prediabetic state, where obesity was induced by a high-fat diet (HFD). Analysis of left ventricular function in the isolated working hearts from HFD-fed mice, suggested that these hearts develop diastolic dysfunction with preserved systolic function. Accordingly, in vivo examination demonstrated maintained systolic function, but we did not find parameters of diastolic function to be altered. In db/db mice, ex vivo working hearts showed both diastolic and systolic dysfunction. Although in vivo functional assessment revealed signs of diastolic dysfunction, the hearts did not display reduced systolic function. The contrasting results between ex vivo and in vivo function could be due to systemic changes that may sustain in vivo function, or a lack of sensitivity using conventional transthoracic echocardiography. Thus, this study demonstrates that the isolated perfused working heart preparation provides unique additional information related to the development of cardiomyopathy, which might otherwise go unnoticed when only using conventional echocardiographic assessment.