In this paper, cross-domain function analysis and trend forecasting in China's construction industry is investigated in order to understand the technical challenges of special construction projects and improve research and development (R&D). The semantic analysis method is used to extract and cluster key context and function information from China's construction patents under Python. Heat maps are then employed to visualize patent function and evolution in different contexts. The most significant function and least significant function are determined using function matrix heat mapping (FMHM), and the logical equations of searching across domain are utilized for lack of specific technology (LST) and lack of common technology (LCT) in cases of technical vacancies. Additionally, time matrix heat map (TMHM) is used to determine if a technology behind a function will be a promising direction or abandoned, providing a macro path for function analysis. The results illustrate that Equipment & Device with functions of ‘Anti-settlement’ and ‘Bar reinforcing’ may be the future direction of Chinese construction, as well as Renovation & Finishing with techniques of ‘Waterproof & Anti-seepage’ and ‘Green energy efficiency’. However, Materials with functions of ‘Masonry’ and ‘Hanging’ may be eliminated due to policy restrictions and technological growth.