This study aims to develop two generalized software reliability growth models (SRGMs) based on the non-homogeneous Poisson process (NHPP), considering the potential for imperfect debugging. The proposed models integrate both exponential and linear characteristics into the fault content function while accounting for varying fault detection rates. Additionally, numerical examples are provided to validate the effectiveness of the proposed models and compare their performance with existing ones. The results indicate that the proposed models offer more accurate prediction capabilities.