It is difficult to reflect the properties of samples from the signal directly collected by the low field nuclear magnetic resonance (NMR) analyzer. People must obtain the relationship between the relaxation time and the original signal amplitude of every relaxation component by inversion algorithm. Consequently, the technology of T2 spectrum inversion is crucial to the application of NMR data. This study optimized the regularization factor selection method and presented the regularization algorithm for inversion of low field NMR relaxation distribution, which is based on the regularization theory of ill-posed inverse problem. The results of numerical simulation experiments by Matlab7.0 showed that this method could effectively analyze and process the NMR relaxation data.