The performances of low dropout linear regulator（LDO） can be lowered to different degrees under the electromagnetic interference

which may affect the whole system’s electromagnetic compatibility. In order to solve this problem

we put forward a modeling method based on BP neural network

and used the genetic algorithm to optimize the initial weights and threshold matrix network. Meanwhile

we used the direct power injection method to design the circuit board for electromagnetic interference injection experiments of LDO within the frequency ranging from 100 MHz to 1 GHz and the power ranging from-15 d Bm W to 25 d Bm W. Taking the output of the LDO as training data

we compared the different structures of the BP neural network prediction performance

and then selected the appropriate network structure. Moreover

we established a model of the LDO electromagnetic interference damage to predict the effects of electromagnetic interference on LDO output data and conducted electromagnetic susceptibility

and made the experimental verification. Finally

we used the model to predict the LDO conducted electromagnetic susceptibility

and compared the predicted data and experimental data of the model. The results show that the maximum relative error between simulation model’s output and the LDO test output is less than 8%

and the maximum relative prediction error between the simulation data of electromagnetic susceptibility of this model and the experiment data is less than 9% in the frequency range of 100 MHz to 1 GHz.