The rapid proliferation of Internet of Things (IoT) devices has produced an unparalleled amount of diverse data, requiring advanced data fusion methods for efficient information extraction and decision-making. This systematic review and meta-analysis evaluate the existing landscape of IoT data fusion methodologies by thoroughly examining literature published from 2020 to 2024 using the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. Out of an initial pool of 112 publications found by systematically searching the Scopus database, 104 articles satisfied the inclusion requirements for systematic review, with 72 articles offering enough quantitative measures for meta-analysis. The research identified notable trends in performance indicators, with F1-scores primarily concentrated in the 95–100% region, signifying the advancement of existing fusion procedures. Dataset sizes showed considerable disparity across several sectors, with security applications using the most extensive datasets (averaging a log-transformed size of 11.7), whilst agricultural applications showed effective performance with smaller datasets (log-transformed size of 6.5). The analysis delineated six primary application domains: security, industrial, smart cities, IoHT (Internet of Healthcare Things), geospatial, and agricultural, each exhibiting unique features and obstacles in implementing data fusion. The results indicate that whereas existing fusion strategies show considerable efficacy across several domains, substantial potential for improvement in resource-limited settings and standardization of assessment criteria persist.