Oil spills in sea water represent one of the most conspicuous forms of damage to the marine environment. They can be detected in SAR images by the exploitation of the normalized cross section (NRCS) and the co-polarization ratio (PR) and by the use of a combined detector. In this paper we describe the simulation of sea surface SAR data by means of the statistical compound model and we use these data for testing the proposed NRCS and PR-based oil spill detection approach. It shows good performance compared to detectors based on the single PR or NRCS information.