For the purpose of investigating the recognition of customer for product attributes, text sentiment score is used to convert online reviews into picture hesitant uncertain linguistic number, and the model of goods attribute analysis is constructed based on the picture hesitant uncertain linguistic generalized Shapely-Choquet weighted averaging operator and online reviews. Firstly, the picture hesitant uncertain linguistic set, the operation of picture hesitant uncertain linguistic number, scoring function and generalized distance measure are defined. The picture hesitant uncertain linguistic generalized Shapely-Choquet weighted averaging operator is proposed, and the related properties of the operator are discussed. Secondly, aiming at the situation that the weight of user group is not completely known, the optimization model is developed to determine the weight of user group. The goods attribute analysis model based on the picture hesitant uncertain linguistic generalized Shapely-Choquet weighted averaging operator and goods online review is proposed. Finally, the model is applied to the attribute analysis of a kind of laptop. The effectiveness of the proposed method is also verified.