The evaluation of left ventricular function is important to identify individuals at high risk for cardiovascular disease. The genetic differences underlying variability in the expression of ventricular function are not totally elucidated. This knowledge will provide a methodology for treatment more selective in retardation myocardial dysfunction. The aim of this study was to identify the relationship between metalloproteinase 3 (MMP3) 5A/6A and metalloproteinase 9 (MMP9) C/T single nucleotide polymorphisms (SNPs) with left ventricular function parameters considered normal and pathologic and to identify the contribution of the polymorphisms on left ventricle ejection fraction (LVEF) variance. Two hundred and five patients underwent myocardial perfusion scintigraphy (MPS) with 99mTc-tetrofosmin, and the left ventricular function parameters were determined. Patients without evidence of ischaemic heart disease which was confirmed with the study of negative MPS were included in Group I and patients with history of ischaemic heart disease which was confirmed with the study of positive MPS to myocardial ischaemia were included in Group II. The MMP3 and MMP9 genotypes were amplified by polymerase chain reaction and were identified by restriction fragment length polymorphism. Individuals carrying the 6A6A genotype (5A/6A MMP3 SNP) had mean values higher left ventricle end systolic volume and left ventricle end diastolic volume and lower LVEF than those carrying the 5A allele (stress: 6A6A: 115 ± 56 ml; 5A allele: 99 ± 44 ml, p = 0.02; 6A6A: 65 ± 51 ml; 5A allele: 49 ± 39 ml, p = 0.04; 6A6A: 49 ± 14%; 5A allele: 55 ± 14%, p = 0.02, respectively), the expression of regional wall thickening and motion abnormalities (p < 0.05) was also more frequent. The frequency of threshold values to predict cardiovascular disease was lesser in heterozygotes than in homozygotes. In Group II, in all the carriers of the 6A6A genotype, abnormal alterations in the LV wall thickening and motion were identified, 17.4% of the carriers of the 5A allele had normal wall thickening and 18.8% had normal wall motion. When the two polymorphisms were included in multiple regression model, the MMP3 5A/6A genetic variant emerged as an independent variable of LVEF (β = − 0.18, p = 0.01). And still, the accumulative effect of two genetic variants (MMP3 5A/6A and MMP9 C/T) had a significant unique contribution to the LVEF variation, in addition to the age and gender (β = 0.14, p = 0.04). The MMP3 5A/6A polymorphism contributes to the variability in the left ventricular function expression. The combined MMP3 5A/6A and MMP9 C/T polymorphisms can be associated with LVEF variance.