In the LTA Formation of Gemah oilfield, northwest Jabung area of Sumatrina Basin of Indonesia, reservoir sand layers are relatively thin. To find out the reservoir sand is the major work to achieve a good oil and gas product; the seismic technology is the key method to identify the distribution of sand layer and also is a big challenge. In this paper, seismic waveform classification and eigenvalue coherency cube have been applied to identify the thin sand layer. Five wells were drilled based on this method, and the outcomes of two wells are consistent with the prediction, which confirms that the integration of multi-seismic attributes is a good method for identifying thin sand layer.