Abstract:Deep formation pore-pressure assessment is very important for evaluation and exploitation in oil and gas fields. Aimed at the deficiency of the present formation pressure assess methods,the evolvement characters of formation pressure time-series is analyzed at first,and then a method is proposed to build the neural network intelligent recognition model. A formation pressure intelligent recognition software is developed based on the MATLAB neural network toolbox. This software is proved to satisfy the engineering needs when used to evaluate the actual formation pore-pressure of well of Q002 in Talimu oil field. In this practice,the recognition average error is 4.02% to normal sediment stratums,and 6.57% to non-normal sediment stratums. This method is simple,efficient,intelligent and can be used to evaluate formation pore-pressure in advance.