In this paper

the Back Propagation（BP）neural network is applied for the purpose of compensating the temperature drift of the pressure sensor and the Davidon Fletcher Powell（DFP）method is adopted to overcome the drawbacks of this network learning algorithm.Algorithm unavailability is effectively solved by search in the descent direction of the L-M learning algorithm and the overflow problem is solved by multiplying a magnification factor added by a disturbance term and thus magnifying the weight updated vector and the weight derivative updated vector.The simulation results show that this method can effectively suppress the effects of temperature on the pressure sensor output and improve the stability and accuracy of the sensor.