Proteomic analysis has the potential to yield vast amounts of data. The available proteomic methods have been hampered by methodological errors in quantification due to large gel-to-gel variations. The inclusion of an internal standard greatly reduces this variation, and therefore the purpose of this investigation was: 1) to develop a sample preparation protocol for human skeletal muscle for two-dimensional differentiated gel electrophoresis (DIGE) and 2) to investigate the repeatability of one particular system, the Ettan DIGE. To test repeatability, nine aliquots from the same homogenate were labelled with three different CyDye(trade mark) dyes (Cy2, Cy3, Cy5). Samples were run on 18 x 24 cm gels, scanned with a Typhoon 9410 laser scanner and analysed in the DeCyder software. When selecting spots appearing only in triplicate (n = 1314), the mean error was 1.7 % (SD: 10.5 %; 95 % CI: 1.1-2.4 %). When setting the significance level to 99 %, no false-positive changes in protein volume ratios were detected. In the protocol presented here, only 0.5 mg tissue was used and separation of >2500 distinct protein spots in the pH range 3-11 and MW 10-200 kDa. Changes in protein abundance of <20 % could be detected. The method is especially useful when comparing muscle proteins between different conditions; for example, healthy and diseased tissue, before and after treatment or different exercise protocols.