Ultra-processed food (UPF) consumption has become a major focus of nutrition research because of its potential associations with adverse health outcomes(1, 2). Accurate assessment of UPF intake is therefore essential; however, the application of the NOVA classification system within food frequency questionnaires (FFQs) remains poorly characterised. Although FFQs are widely used to assess habitual dietary intake, variations in food item detail, classification procedures, and reporting practices may affect the accuracy, reproducibility and comparability of UPF estimates. This systematic review aimed to evaluate how the NOVA classification system has been applied within FFQs used to assess UPF consumption and to identify factors influencing the suitability of FFQs for processing-based food classification. The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (3) and prospectively registered with PROSPERO (CRD420261382629). Searches were undertaken in MEDLINE (Ovid), PubMed, Scopus, Embase and Web of Science. Eligible studies assessed dietary intake using FFQs and applied the NOVA classification system to estimate UPF consumption. EndNote 21 was used for reference management and duplicate removal, and Covidence was used for title, abstract and full-text screening. Data extraction was conducted using standardised forms capturing study characteristics, FFQ methodology, and NOVA classification procedures. To evaluate methodological quality, a novel NOVA–FFQ scoring framework was developed. The framework assesses three domains: distinction between branded and non-branded or preparation sources; details relating to ingredients and processing attributes; and validation or adaptation for processing-based classification. Scores range from 0 to 23, allowing systematic comparison of methodological strengths and limitations across FFQs. The search identified 3970 records, of which 484 remained after title and abstract screening. Full-text screening and data extraction are ongoing. Preliminary synthesis has identified recurring challenges, including assumptions made during NOVA classification, classification of composite foods, and limited reporting of classification procedures. Preliminary evaluation of four FFQs demonstrated substantial variation in methodological suitability, with scores ranging from 7 to 15 out of 23. The MANS FFQ achieved the highest score (15/23), followed by the Harvard FFQ (10/23), Elsa-Brasil FFQ (9/23) and Block FFQ (7/23). These findings suggest that retrospective application of NOVA categories to FFQ-derived dietary data may compromise consistency and reproducibility. This review will provide the first comprehensive evaluation of NOVA application within FFQs and establish a structured approach for assessing FFQ suitability for UPF research.