This paper compares the classification capability of data acquired in hybrid and dual linear polarization mode over the Chelmsford area, United Kingdom, from RISAT-1 C-band satellite. Support vector machine based supervised classification is used in the study and accuracy is assessed over the validation pixels. The hybrid-pol RH/RV combination shows better classification accuracy over linear pol HH/HV combination by 2.5 %. Overall classification accuracy is increased to 90% over the given area when Stokes parameter are added to the polarization combination.