A　formalized　model　for　service　compositions　is　established　based　on　the　Markov　chain.　It　is　called　the　Markov　chain　model.　The　model　can　be　established　by　translating　the　definition　specifications　to　the　Markov　chain　when　the　definition　specifications　of　composite　services　are　given,　or　by　analyzing　the　executing　history　data　when　the　definition　specifications　of　composite　services　are　unknown.　Different　from　most　of　other　formalized　models　with　the　only　function　of　verifying　composite　services＇　correctness,　the　Markov　chain　model　for　composite　services　have　many　uses,　such　as　evaluating　the　trustiness　of　executing　composite　services,　getting　composite　services　executing　costs,　for　example,　response　time,　expenses,　etc.　The　most　important　one　is　that　most　of　application　algorithms　for　the　Markov　chain　model　can　effectively　solve　the　polynomial　complexity,　while　other　formalized　models＇　majority　of　problems,　such　as　the　verification　problem,　are　NP-hard　problems.