Abstract Aims Left ventricular function recovery (LV-REC) or left ventricular adverse remodelling (LV-REM) after acute myocardial infarction (AMI) play an important role for identifying patients at risk of heart failure. In this study we aim to evaluate the usefulness of non-invasive myocardial work (MW), a new index of global and regional myocardial performance, to predict LV-REC or LV-REM after AMI. Methods and results Fifty patients with AMI (mean age, 63.8 ± 13.4 years), treated by primary percutaneous coronary intervention (PCI), were prospectively enrolled. They underwent a baseline transthoracic Doppler echocardiography (TTE) within 48 h after PCI and a second TTE after a median of 31 days during the follow-up. MW was derived from the strain-pressure loops, integrating in its calculation the non-invasive arterial pressure, according to standard speckle tracking echocardiography recommendations. LV-REC was defined as an absolute improvement of left ventricular ejection fraction (LVEF) ≥ 5% from LVEF at baseline, whereas LV-REM was defined as an increase of ≥ 20% of the LV end diastolic volume (LVEDV) at 1 month follow-up. We overall found a significant improvement from baseline to one-month follow-up for values of LVEF (49.8 ± 9.5% vs. 52.8 ± 9.3%, P = 0.001), global longitudinal strain (GLS) (−13.4 ± 3.9% vs. −18.7 ± 5.4%, P = 0.016), global work index (GWI) (1368.6 ± 435.2 vs. 1788 ± 493 mmHg/%, P = 0.0001), global work efficiency (GWE) (89.96 ± 9.3% vs. 91.3 ± 6.4%, P = 0.001), global constructive work (GCW) (1619.16 ± 497.9 mmHg/% vs. 2008.6 ± 535.3 mmHg/%, P = 0.0001), global wasted work (GWW) (188.8 ± 19.8 mmHg/% vs. 149.2 ± 16.5 mmHg/%). However, LV-REC at 1 month of follow-up was observed only in 36% of the population enrolled, whereas LV-REM was described in 18% of cases. Using ROC curve analysis, we identified a cut off value of 202 mmHg/% for baseline GWW (sensitivity 75%, specificity 62%, AUC 0.6667, CI 95%: 0.51618–0.81715, P = 0.0001) to identify patients with LV-REM at 1 month. With regards to conventional echo parameters, patients with LV-REC showed lower baseline wall motion score index (WMSI) than those without LV-REC (1.73 vs. 1.38, P = 0.007). Conclusions Among standard and advanced TTE parameters, only baseline GWW is able to predict early LV-REM at 1 month after primary PCI. Therefore, it could be used during baseline evaluation of AMI patients for a more accurate stratification of those at higher risk of heart failure. However, further larger scale studies are needed to validate these findings.