Traditional scene-based nonuniformity correction algorithms for infrared focal plane array suffer from two major drawbacks: slow convergence rate and low correction accuracy. Aiming at these problems, a new method based on extended total variation is proposed to correct the nonuniformity for infrared focal plane arrays. Based on the analysis of total variation-based de-noising performance, the scope of total variation is extended for motional infrared image sequences. By minimizing the total variation of corrected images, iterative formula to compute the gain and offset factors is obtained using steepest descent method. For eliminating the ghosting effect, an adaptive threshold method is designed. The experiments show that compared with the existing methods, this method can effectively remove the noise in original infrared images and largely retain image information at the same time. The image quality is therefore improved.