<p>Innovation is a process of "descent with modification" that gradually changes based on previous innovations. Therefore, deriving trajectories connecting past and present innovations is important for analyzing the dynamic process of innovation and predicting future innovations. This study aimed to derive the citation relationship connecting past and present innovations for products and presents it as a model called a product citation network. To derive the citation relationship between products, we employed embedding and clustering techniques to quantify individual products and measure the similarity between products from different time periods. We conducted an empirical study on mobile product data released in the global market from 1994 to 2023 in order to validate the proposed methodology. We applied the disruption index, which is widely used in recent innovation studies, to analyze the product innovation patterns of each firm and thus verify that the constructed mobile product citation network has sufficient explanatory power for actual mobile product innovation cases. This study demonstrates the applicability of the innovation studies approach using paper and patent citation networks in product innovation studies and provides a basis for advancing product innovation research aligning with the latest advances in innovation studies.</p>