The invention discloses a non-intrusive load decomposition method based on a sparse classifier hierarchical algorithm. The method comprises the following steps: S1, acquiring voltage and current dataof an electric appliance; s2, calculating low-frequency active power, low-frequency reactive power and high-frequency current harmonics of the electric appliance; s3, performing clustering analysis; s4, constructing a feature dictionary; s5, constructing a sparse matrix; s6, performing power identification; s7, judging a result, and outputting a judgment result if the overlapping number in the judgment result is equal to 1; otherwise, executing the step S8; and S8, performing current harmonic identification and outputting an identification result. By adopting the non-intrusive load decomposition method based on the sparse classifier hierarchical algorithm, whether current harmonic identification is introduced or not can be determined by comprehensively considering active power identification and reactive power identification results of the electrical appliances when part of the electrical appliances with similar power exist in users, so it is ensured that when the power is similar, long-time and large-calculation-amount electric appliance identification does not need to be performed, and the electric appliance identification result can be guaranteed.