Abstract: Firstly, this study constructs a three-dimensional technology foresight framework integrating "technological value, structural relatedness, and examination timeliness" by developing a multi-dimensional patent value evaluation system, employing the entropy-weighted TOPSIS method and multiple linear regression models to derive comprehensive patent value scores and influencing factors. Second, patent datasets are categorized based on examination speed, and BERTopic model are applied to identify the technological themes, enabling a comparative analysis of their thematic distributions. Finally, a three-dimensional technology foresight mapping framework is developed, incorporating theme value, theme relevance, and average examination duration to delineate the technological prospects embedded in fast and slow patents. Taking the smart grid sector as an example, the findings reveal that within the dimension space defined by high thematic relevance and high thematic value, slow patents demonstrate stronger technological prospect potential compared to fast patents. Based on these insights, strategic recommendations are proposed to assist senior R&D managers in effectively advancing innovation.

Key words: Fast and slow patent analysis, Technology foresight exploration, Three-dimensional technology map, Strategic emerging industry, BERTopic model