Abstract: Prestacking simultaneous inversion provide a tool to fully use prestacking seismic information. In combination with production data, it is possible to reveal changing lithology of and liquid in a reservoir. Of course, there are uncertainties in data interpretation. The uncertainties could be estimated if Bayes criterion is used in interpretation of prestacking simultaneous inversion data. Rock physics simulation, prestacking inversion and probability density function construction remain the keys of the technique. Data of well logging and rock physics simulation are the information required prior to the processing,and inversion data can be used as the basic information to estimate the probability density function of different lithofacies. This technique takes all the advantages of using probability and statistic as tools to quantitative interpret prestacking inversion data, and it is more reliable for complex geological targets with multi-solution, such as characterization and identification of lithofacies. As to the application of this technique to reservoir prediction in igneous rocks, it shows that Bayes Criterion is much better to identify gas bearing reservoirs, and thus could better serve the prediction of gas-bearing igneous rocks.