Abstract: In the forecast of reservoir lithology with the impedence invension and conventional seismic attribute techniques,it is very difficult to predict the thin reservoirs veraciously for lack of drilling wells.The Stratimagic waveform analysis technique makes up this shortage of predicting reservoirs with the conventional methods.By taking full advantage of the abundant information in the seismic data,this technique classifies the seismic traces of target layers with the neural network algorithm,describes the horizontal changes of seismic signals,and obtains the maps of seismic facies matching to the geological facies.This technique can be used to predict the distribution of reservoir sand beds and subtle oil reservoirs. Taking the lithologic oil reservoir of Bamianhe oilfield in Dongying depression as a representative example,this paper introduces the application of this technique.Favorable results are achieved by this technique in this area,and several well locations provided by this method produced oil successfully,showing that this technique is effective to predict the reservoir sand beds and the subtle oil reservoirs.