Abstract: Targets can be detected and identified by the coherence coefficient criteria of the scattering vector of ideal target and to be classified. Polarimetric scattering is computed by the software BART developed by Key Laboratory for Information Science of electromagnetic Waves, scattering vector of ideal target is then constructed, and synthetic aperture radar(SAR) images are also simulated. Based on the simulated data as sample and measured Pi-SAR image of cultivation boxes over sea surface, detection and identification algorithm by coherence coefficient criteria is studied. Results show that this algorithm is an effective target detection approach.