Abstract:I believe that the magnetic anomaly due to long-period hydrocarbon seeping from hydrocarbon-bearing sedimentary formation is a secondary magnetized field.Such secondary magnetized field can be discovered by interpreting accurate aeromag-netic data processed with neural network method. In the method,magnetic anomaly intensity,magnetic anomaly width,shape ratio of magnetic anomalies before and af-ter polarization,and magnetic anomaly gradient are taken as control variates which characterize magnetic field. We did pattern recognition processing of aeromagnetic data over PFW oil field,with 87% success.