The invention relates to a method and an apparatus for optimizing a human face picture quality evaluation model. The method comprises the steps of establishing a human face picture test set; identifying the similarity between a to-be-tested human face picture and a sample human face picture in a preset human face database, and according to the similarity and picture identity information, obtainingan identification result of each to-be-tested human face picture; according to the identification result, determining a quality score of each to-be-tested human face picture; and performing neural network training by taking the to-be-tested human face pictures and the corresponding quality scores as training data to obtain optimized human face picture quality evaluation model and parameters. Through the human face picture quality evaluation model and the parameters, human face picture quality evaluation is not influenced by human subjective factors.