Abstract: Corrosion rate is directly related to safe and effective service of biodegradable Zn alloys after implantation. Currently, no highly accurate and interpretable “white-box” machine learning models exist for predicting corrosion rates. This work proposes a data-driven accelerated corrosion method for predicting corrosion rates of biodegradable Zn-0.45Mn-0.2Mg alloys. A symbolic regression (SR) machine learning model is established for the first time based on an analytical expression among corrosion rate and four corrosion parameters. The SR model with determination coefficient of 0.97 outperforms five other machine learning models. Prediction errors of verification experiments are all below 10%. This work marks a paradigm shift in corrosion researches of biodegradable metals from qualitative to quantitative analysis.