Proteases play a crucial role in the soil nitrogen (N) cycle by converting protein to oligopeptides and amino acids. They are often viewed as a bottleneck in terrestrial N cycling; therefore, it is vital that we have robust methods for evaluating protease activity in soil to understand global patterns of protease activity. In response to this, several laboratory-based protease methods have been developed and subsequently modified. However, the validity of these different approaches remains largely unknown. In addition, the lack of standardised protocols makes it difficult to compare protease activity across studies. In this systematic synthesis, we critically evaluate the most common colorimetric and fluorimetric methods used to measure soil protease activity involving 680 independent studies and 1,491 individual assays. To investigate the key regulators of soil protease activity, we collected associated metadata on environmental (mean annual temperature and soil pH) and methodological (assay temperature and pH) factors. Protease activity measured with colorimetric substrates were centred around ca. 1000 nmol product g−1 h−1, whilst rates measured with fluorimetric substrates were lower at ca. 100 nmol product g−1 h−1. Fluorimetric and colorimetric substrates target different proteases which are likely to have different abundances, kinetic parameters, catalytic mechanism or ecological function suggesting why colorimetric substrates have a higher protease activity. We found soil protease activity varied widely around these peaks, likely due to a wide range of environmental or methodological factors that may influence/bias the result. We present the following recommendations for measuring soil protease activity: 1) report assay conditions and soil characteristics, particularly pH and temperature, 2) conduct the assay at either field or optimised pH and temperature conditions, and, 3) check that measurements lie between 0 and 5000 nmol product g−1 h−1. This will help reduce the variation in soil protease activity measurements due to methodological bias and improve reporting of abiotic and biotic associated data. Altogether this will lead to a better understanding of the ecological drivers of protease activity and refine parameterisation of global biogeochemical models.