Soil DNA extraction has become a critical step in describing microbial biodiversity. Historically, ascertaining overarching microbial ecological theories has been hindered as independent studies have used numerous custom and commercial DNA extraction procedures. For that reason, a standardized soil DNA extraction method (ISO-11063) was previously published. However, although this ISO method is suited for molecular tools such as quantitative PCR and community fingerprinting techniques, it has only been optimized for examining soil bacteria. Therefore, the aim of this study was to assess an appropriate soil DNA extraction procedure for examining bacterial, archaeal and fungal diversity in soils of contrasting land-use and physico-chemical properties. Three different procedures were tested: the ISO-11063 standard; a custom procedure (GnS-GII); and a modified ISO procedure (ISOm) which includes a different mechanical lysis step (a FastPrep ®-24 lysis step instead of the recommended bead-beating). The efficacy of each method was first assessed by estimating microbial biomass through total DNA quantification. Then, the abundances and community structure of bacteria, archaea and fungi were determined using real-time PCR and terminal restriction fragment length polymorphism approaches. Results showed that DNA yield was improved with the GnS-GII and ISOm procedures, and fungal community patterns were found to be strongly dependent on the extraction method. The main methodological factor responsible for differences between extraction procedure efficiencies was found to be the soil homogenization step. For integrative studies which aim to examine bacteria, archaea and fungi simultaneously, the ISOm procedure results in higher DNA recovery and better represents microbial communities.