During the 14th Five-Year Plan period (2021–2025), technological innovation has become a key driver of the high-quality development of China's publishing industry. With the rapid advancement of artificial intelligence (AI), big data, and other emerging digital technologies, innovation activities have increasingly permeated all segments of the industry. However, limited attention has been paid to the structural distribution of innovation activities across the publishing industry chain and their evolutionary trajectories based on objective innovation output data. With this in mind, this study applies the Innovation Value Chain (IVC) framework to examine the structural characteristics and development trends of technological innovation in China's publishing industry. Drawing on 28 A-share listed publishing companies as the sample, this study collected patent data from the China National Intellectual Property Administration (CNIPA) database and software copyright registration data from enterprise information platforms. A total of 889 patents and 2, 148 software copyright registrations filed between 2021 and 2025 were analyzed. Through bibliometric analysis and BERTopic-based topic modeling, the study investigates the scale, thematic structure, and distribution of innovation outputs across the innovation value chain. Innovation themes were further mapped onto upstream, midstream, and downstream segments of the publishing industry chain to reveal the allocation of innovation resources. The findings indicate that technological innovation in the publishing industry exhibits a markedly uneven structure. Innovation outputs are concentrated among a small number of leading firms and display evident regional agglomeration effects. Software copyright registrations substantially outnumber patents, suggesting that publishing enterprises place greater emphasis on platform construction, application development, and service innovation. Innovation activities are mainly concentrated in the midstream production stage and downstream application stage. Patent-based innovations primarily focus on production process optimization, digital reading technologies, equipment improvement, and user-oriented applications, whereas software-based innovations are largely devoted to digital platforms, copyright management, content creation support, and intellectual property (IP) commercialization. By contrast, upstream innovation related to original content technologies, copyright governance, and data infrastructure remains relatively limited. Consequently, the current innovation landscape is dominated by short-chain and medium-chain innovation, while long-chain innovation remains comparatively underdeveloped. The study further suggests that technological innovation in the publishing industry is undergoing a transition from production-oriented technological improvement toward application-oriented service expansion. Innovation activities are increasingly moving from technology adaptation to publishing-specific technological development, from isolated process optimization to full-process digital collaboration, and from single-content monetization to platform-based and multi-scenario value realization. Platform-based development, AI-enabled transformation, and data-driven innovation are emerging as major development trends. By revealing the structural distribution and evolutionary trends of technological innovation, this study enriches the application of the Innovation Value Chain theory in publishing research and provides empirical evidence for optimizing innovation allocation, improving innovation quality, and fostering new-quality productive forces in the publishing industry.