Crambe abyssinica is a non-edible crop able to adapt to adverse climatic conditions, arising as an alternative feedstock for biodiesel production. With this objective, an eco-friendly route for biodiesel production from crambe oil by enzymatic transesterification was evaluated in order to maximize fatty acid methyl esters content. Response Surface Methodology and central composite design (two factors, three levels, one central point, two replicas) was used to obtain predictive models considering catalyst concentration and methanol:oil molar ratio as independent variables. Since the reaction was monitored overtime, a predictive model considering three independent variables (three factors, three levels) was also obtained. Enzymatic transesterification is a good alternative process to obtain high FAME yield, with 84 ± 5 wt % being obtained after 2 h using a methanol:oil molar ratio 6:1 and 8 wt% of enzyme, a value which predictably can increase to almost 99 wt% after 24 h. The results showed that the methanol:oil molar ratio and the enzyme’s concentration considerably influence the FAME content during enzymatic catalyzed reaction. The polynomial regression models obtained can help to balance the concentration of enzyme and alcohol to improve FAME production and to predict the resulting yield.