Despite recent technological advances in molecular ecology, DNA extraction from soil remains a crucial step when quantifying and characterizing soil microbial communities. Potential biases could hamper fundamental understanding of the dynamic relationships between soil properties and microorganisms under different agricultural practices. In this study, we compared four different DNA extraction methods in their ability to discriminate microbial communities of an arable soil subjected to different fertilization managements for more than 50 years (Ultuna, Sweden). The abundance and the diversity of bacteria, archaea, and fungi were studied using qPCR and molecular fingerprints, respectively. Overall, the choice of DNA extraction method had a more pronounced effect on the fungal and archaeal communities in comparison to bacterial ones. The assessment of the microbial diversity was more sensitive to DNA extraction methods in comparison with the quantification of the abundances. The DNA extraction method clearly affect the intensity of the correlations between the abundance and/or diversity of microbial communities and environmental variables (C, N and pH) according to the targeted taxon. This study highlights that long-term effects can be offset by biases in DNA extraction methods.