In the present work multivariate regression models using interval partial least square (iPLS) and backward interval partial least square (biPLS) had been analyzed and compared. iPLS and biPLS models had been developed to determine the concentration of biodiesel in blends of biodiesel/diesel using infrared spectroscopy signals. 45 samples with concentrations in range 8-30% of biodiesel, and two distinct spectrophotometers were used. Both the techniques (iPLS and biPLS) using the data obtained by HATR-FTIR if had shown promising to develop simpler, faster and non-destructive methodologies for the biodiesel determination in commercial blends.