With the growth of the commodity economy, packaging has evolved beyond its original function of protecting goods to play a crucial role in capturing consumer attention, influencing purchasing decisions, and building brand identity. From a consumer perspective, emotional design in packaging addresses both their emotional aspirations and desire for a quality life. This study focuses on the emotional design of product packaging, aiming to meet consumers’ material needs while also fulfilling their deeper, spiritual expectations, ultimately enhancing the consumer experience and strengthening brand connection. This topic is based on MEC(Mobile Edge Computing), the research of commodity packaging design and emotion perception recognition. Design an algorithm framework of trademark pattern integrity detection and a model for image emotion classification and recognition. The binary image of trademark pattern is obtained by color image segmentation method, then the shape feature descriptor of trademark pattern is calculated, and the feature dimension is reduced by PCA(principal component analysis) method to reduce irrelevant features and feature data redundancy. The results show that the performance of this method in accuracy, recall, and F1 value are improved by 7.1%, 19.02%, and 11.20%, respectively. It provides a theoretical basis for how to choose the right starting point for packaging design and broadens the horizon for the future research of packaging design.