Intact proteins comprise a major component of organic carbon and nitrogen produced globally and are likely an important fraction of organic matter in sediments and soils. Extracting the protein component from sediments and soils for mass spectral characterization and identification represents a substantial challenge given the range of products and functionalities present in the complex matrix. Multiple forms of gel electrophoresis were evaluated as a means of enhancing recovery of sedimentary protein before proteomic characterization and compared with a direct enzymatic digestion of proteins in sediments. Resulting tryptic peptides were analyzed using shotgun proteomics and tandem mass spectra were evaluated with SEQUEST. Multiple databases were then evaluated to examine the ability to confidently identify proteins from environmental samples. Following evaluation of electrophoretic extraction of proteins from sediments, the recovery of an experimentally added standard protein (BSA) from older (>1 ky) sediments was optimized. Protein extraction from sediments via direct electrophoresis of a slurry mixture and the specified extraction buffer resulted in the greatest number of confident protein identifications and highest sequence coverage of the BSA standard. Searching tandem mass spectral data against larger databases with a higher diversity of proteomes did not yield a greater number of, or more confidence in, protein identifications. Regardless of the protein database used, identified peptides correlated to proteins with the same function across taxa. This suggests that while determining taxonomic‐level information remains a challenge in samples with unknown mixed species, it is possible to confidently assign the function of the identified protein.