Due to being influenced by detectors' material,related manufacturing technology etc,every detection element's responsivity of infrared focal plane arrays(IRFPA) is different,which results in non-uniformity of IRFPA.So non-uniformity correction(NUC) is an important technique for IRFPA.The traditional NUC algorithms based on neural network were analyzed in this study.And a new improved NUC algorithm based on neural network was presented. A comparison was made to these algorithms.Experiments results show that the algorithm proposed in this study prevails over other algorithms based on neural network in correction effective and convergence speed.