The ubiquity and transferability of soil makes it a resource for the forensic investigator, as it can provide a link between agents and scenes. However, the information contained in soils, such as chemical compounds, physical particles or biological entities, is seldom used in forensic investigations; due mainly to the associated costs, lack of available expertise, and the lack of soil databases. The microbial DNA in soil is relatively easy to access and analyse, having thus the potential to provide a powerful means for discriminating soil samples or linking them to a common origin. We compared the effectiveness and reliability of multiple methods and genes for bacterial characterisation in the differentiation of soil samples: ribosomal intergenic spacer analysis (RISA), terminal restriction fragment length polymorphism (TRFLP) of the rpoB gene, and five methods using the 16S rRNA gene: phylogenetic microarrays, TRFLP, and high throughput sequencing with Roche 454, Illumina MiSeq and IonTorrent PGM platforms. All these methods were also compared to long-chain hydrocarbons (n-alkanes) and fatty alcohol profiling of the same soil samples. RISA, 16S TRFLP and MiSeq performed best, reliably and significantly discriminating between adjacent, similar soil types. As TRFLP employs the same capillary electrophoresis equipment and procedures used to analyse human DNA, it is readily available for use in most forensic laboratories. TRFLP was optimized for forensic usage in five parameters: choice of primer pair, fluorescent tagging, concentrating DNA after digestion, number of PCR amplifications per sample and number of capillary electrophoresis runs per PCR amplification. This study shows that molecular microbial ecology methodologies are robust in discriminating between soil samples, illustrating their potential usage as an evaluative forensic tool.