Background Metformin-associated lactic acidosis (MALA) is a rare event but underrecognition may lead to unfavorable outcomes in type 2 diabetes patients. While many risk factors of MALA have been identified, how to reduce mortality from MALA is a matter of debate. This study aimed to explore the factors associated with 30-day mortality amongst MALA patients. Methods An observational study enrolled patients diagnosed with MALA between January 2014 and December 2017. MALA was defined by a history of metformin administration, metabolic acidosis (arterial blood gas pH <7.35 or HCO3 <15 mmol/L), and elevated plasma lactate level (>5 mmol/L). We examined risk factors including age, sex, underlying diseases, current medications, blood tests, disease severity, and dialysis data. Mortality status was identified from medical records or report on telephone. Results We included 105 MALA patients. Most patients (95.2%) were diagnosed acute kidney injury stage 3 according to KDIGO 2012 definition. The 30-day mortality rate was 36.2% and dialysis rate was 85.7%. The survivors had higher proportions of underlying chronic kidney disease, presence of metabolic acidosis, receiving renal replacement therapy within 6 hours, and haemodialysis, whereas the non-survivors had higher percentage of hypertension and disease severity. Lower APACHE II score (HR = 0.95; 95%CI, 0.91–0.99; p = 0.038), time to dialysis < 6 hours (0.31; 0.14–0.69; 0.004), and haemodialysis (0.20;0.06–0.67; 0.010) were associated with lower 30-day mortality, using multivariate Cox-regression analysis. Conclusions Mortality rate amongst patients with MALA was high. Early dialysis treatment within 6 hours after admission and haemodialysis were independently associated with lower 30-day mortality. The large scale, well-designed studies need to confirm these encouraging results.