Aiming at the lack of a complete and comprehensive decomposition method for independent load in the existing non-intrusive residential electricity load monitoring, the integrity of electricity consumption information cannot be guaranteed. An improved blind source separation algorithm based on ensemble empirical mode decomposition (EEMD) and Pearson-PCA is proposed. Firstly, EEMD is used to decompose the total power signal to eliminate the modal aliasing phenomenon in the empirical mode decomposition process, and it can obtain a series of intrinsic mode functions (IMF). Secondly, combining with Pearson correlation coefficient and principal component analysis (PCA), an improved Pearson-PCA algorithm is proposed to reduce the dimensionality of the IMF, remove the weaker IMF components, and estimate the number of source signals. Then, fast independent component analysis (FastICA) is used to decompose the reduced-dimensional IMF to calculate the source power signal. Finally, the proposed improved algorithm is applied to the non-intrusive residential electricity load decomposition problem, and the reference energy disaggregation data (REDD) is used for experimental simulation. The experimental results show that the proposed improved algorithm has a better decomposition effect in different electricity consumption scenarios.