The escalating fossil fuel costs and over-dependency on imports are driving the research in the domain of biodiesel. The increasing greenhouse gas emissions are further driving the investigation in this domain. In the present study, an engine is tested using waste-derived biodiesel. The focus was on engine emission, and how the use of biodiesel influences it. The data concerning the four emission species oxides of nitrogen (NOx), unburnt hydrocarbon (UHC), carbon monoxide (CO), and carbon dioxide (CO2) were collected. Since the formation of emissions in a compression ignition engine is a complex process, it is difficult to model using conventional approaches. Hence, an evolutionary machine learning technique Gene Expression Programming (GEP) was employed to model and predict the emission data in this study. The statistical evaluation of the data revealed that GEP could predict the data with more than 95.64% efficacy both during model training as well as model training phase. The model was also tested on different statistical metrics like mean squared error and mean absolute error. All these statistical metrics showed that GEP performed well throughout the model development.