In Bayan Hot basin of northern China, igneous rocks are characterized by small varying range, small thickness and deep burial depth. Their gravity and magnetic anomalies are weak, and they do not have remarkable resistivity anomalies. They fail to form continuous and strong lineup and hence cannot be distinguished by a single geophysical prospecting method. With high precision gravity and magnetic data and magnetotelluric sounding data, however, we can determine their locations and burial depths. Using logging data of X1 log, the authors made synthetic seismic record to analyze the seismic wave characteristics of the igneous rocks, and then set up the training swatches of BP Neural Network to analyze the distribution of the igneous rocks. It is proved that the synthetic geophysical interpretation is an effective method for improving the geophysical inversion and interpretation.