This study is designed to test the agreement in measuring left ventricular systolic function between transthoracic two-dimensional echocardiography (2DTTE) and "gold standard" of non-invasive cardiac imaging, magnetic resonance imaging (CMRI) and their impacts on the classification of patients according to the left ventricular ejection fraction (EF). 32 patients who were suspected with heart disease were evaluated by CMRI and 2DTTE examinations. End diastolic volume (EDV), end systolic volume (ESV), EF and left ventricular function category were then calculated and compared. There was no significant difference (P=0.504) for EDV, while ESV of CMRI was significantly higher than that of 2DTTE (P=0.049), and EF of CMRI was significantly lower than that of 2DTTE (P= 0.018). There was no significant difference (P=0.077) in left ventricular functional category. Bland-Altman analysis of LV volumetric data and EF measurements showed a good agreement between two methods. The 2DTTE over-estimated I (n=5) or II (n=1) degrees of functional classification when compared with the CMRI. Both CMRI and 2DTTE are of great clinical value in evaluating left ventricular systolic function, while CMR may be more beneficial to patients with abnormal LV functions.