Sentiment classification is also called tendency analysis, which is a text classification task in the field of natural language processing. The object of traditional text classification tasks is the objective content of the text, while the object of tendency analysis is the "subjective factors" of the text. Aspect-Based sentiment classification is a fine-grained classification task, whose purpose is to judge the sentiment polarity of the aspect given in the text. The article mainly summarizes the aspect-based sentiment classification method. First, sort out and classify the aspect-level sentiment classification methods based on deep learning at home and abroad, and subdivide them into convolutional neural network methods, memory neural network methods, recurrent neural network methods, recurrent neural network methods, and attention recurrent neural network methods, and the method based on graph convolutional network and attention graph convolutional network. Then, the research work is introduced by category, focusing on the method based on graph convolutional network and attention graph convolutional network as the latest research progress. Secondly, it briefly introduces the specific application scenarios of sentiment classification, and organizes the commonly used data sets of aspect-based sentiment classification for researchers' reference. Finally, the development of aspect-based emotion classification is summarized and prospected. This study provides a comprehensive review reference for researchers in the field of aspect-based sentiment classification.