Abstract：The distributions of reservoir parameters are ascertained usually by using seismic data or logging data respectively, but they are analysed rarely by adopting both seismic data and logging data simultaneously. Here reservoir parameters are predicted by taking Bayse-Kriging estimation technique with simultaneous use of seismic and logging data. In this technique, linear Bayes theory is applied in Kriging estimation. It is achieved by first constructing a model to class the data needed in estimation into two types(observed data and guess data), then using areal variable theory to analyse the spacial variation of the two type data. As to practical seismic exploration and logging, logging data may be taken as observed data, and seismic data as guess data. The real prediction of thickness and porosity of sandstone in Wangju-Caojiawu area proves this technique feasible.
