New quality productivity emphasizes breaking through the limitations of traditional productivity development models and achieving significant improvements in total factor productivity through innovative ways of allocating production factors. In this context, artificial intelligence (AI), as an emerging driving force, is continuously injecting strong impetus into Guangdong’s economic development and industrial structure (IS) transformation and upgrading. The government has implemented a series of supportive policies to accelerate the deep integration of AI with the three major industries, promote the transformation of economic development mode, and optimize and upgrade IS. This article proposes a Guangdong IS optimization and resource allocation path model that integrates deep neural networks (DNN) and quantitative analysis based on new quality productivity, aiming to further promote the upgrading of Guangdong’s digital industry. The results indicate that the model can effectively improve resource allocation efficiency, promote IS optimization, and provide new path choices for the high-quality development of Guangdong’s economy. This study not only provides theoretical support for the upgrading of IS in Guangdong but also provides a useful reference for the economic development of other regions in the background of new quality productivity.