Identifying sustainable agricultural practices to support policy development requires a rigorous synthesis of scientific evidence based on experiments carried out around the world. In agricultural science, meta-analyses (MAs) are now commonly used to assess the impact of farming practices on a variety of outcomes, including crop and livestock productions, biodiversity, greenhouse gas emissions, nitrate leaching, soil organic carbon, based on a large number of experimental data. MA has become a gold standard method for quantitative research synthesis, and the growing number of MAs available can potentially be used to inform decisions of policy makers. However, published MAs are heterogeneous both in content and quality, and a framework is needed to help scientists to report the results and quality levels of MAs in a rigorous and transparent manner. Such a framework must be implementable quickly - within weeks - to be operational and compatible with the time constraints of modern policymaking processes. In this paper, we propose a methodological framework for assessing the impacts of farming practices based on a systematic review of published MAs. The framework includes four main steps: (1) literature search of existing MAs, (2) screening and selection of MAs, (3) data extraction and quality assessment, and (4) reporting. Three types of reports are generated from the extracted data: individual reports summarizing the contents of each MA (MA summary reports), reports summarizing each of the impacts of a given farming practice on a specific environmental, climate mitigation, or production outcome (single-impact reports), and report summarizing all the impacts of a given farming practice on all the outcomes considered (general report). All these reports present the quality levels of the MAs examined on the basis of 16 quality criteria. The proposed framework is semi-automatic in the sense that the skeletons of the reports are generated automatically from the spreadsheet used for the data extraction and quality assessment. This semi-automatic procedure allows scientific experts to reduce the time needed in the reporting step. Since 2020, the proposed framework was successfully applied by a group of scientific experts to support decisions of EU policy makers, and examine a large diversity of single farming practices (e.g. nitrification inhibitors, biochar, liming) and cropping systems (e.g. organic systems, agroforestry) in a relatively short period of time. It provides an operational tool for scientists who want to supply policymakers with scientific evidence based on large numbers of experiments, in a timely and reproducible manner.