Metaproteomics and its potential applications are very promising to study microbial activity in environmental samples and to obtain a deeper understanding of microbial interactions. However, due to the complexity of soil samples the exhaustive extraction of proteins is a major challenge. We compared soil protein extraction protocols in terms of their protein extraction efficiency for two different soil types. Four different protein extraction procedures were applied based on (a) SDS extraction without phenol, (b) NaOH and subsequent phenol extraction, (c) SDS–phenol extraction and (d) SDS–phenol extraction with prior washing steps. To assess the suitability of these methods for the functional analysis of the soil metaproteome, they were applied to a potting soil high in organic matter and a forest soil. Proteins were analyzed by two-dimensional liquid chromatography/tandem mass spectrometry (2D-LC–MS/MS) and the number of unique spectra as well as the number of assigned proteins for each of the respective protocols was compared. In both soil types, extraction with SDS–phenol (c) resulted in “high” numbers of proteins. Moreover, a spiking experiment was conducted to evaluate protein recovery. To this end sterilized forest soil was amended with proteins from pure cultures of Pectobacterium carotovorum and Aspergillus nidulans. The protein recovery in the spiking experiment was almost 50%. Our study demonstrates that a critical evaluation of the extraction protocol is crucial for the quality of the metaproteomics data, especially in highly complex samples like natural soils.