In　the　cloud　manufacturing　(CMfg)　platform,　there　are　numerous　manufacturing　services　with　same　or　similar　functions.　Due　to　the　inconsistency　of　service　quality,　how　to　effectively　evaluate　the　quality　of　services　is　a　fundamental　problem　in　CMfg,　and　which　aims　to　reduce　risk　and　increase　the　benefits　of　users.　Under　these　contexts,　this　study　first　constructs　a　comprehensive　three-dimensional　trust　evaluation　system　that　considers　the　trust　of　service　demanders,　resource　providers,　and　cloud　platform　operators　in　CMfg.　And　then,　a　manufacturing　services　trust　evaluation　and　preprocessing　model　is　proposed,　and　the　optimization　process　is　described　as　following:　(1)　superior　and　inferior　manufacturing　services　are　identified　by　the　first　stage　filtration;　and　(2)　inferior　manufacturing　services　are　further　classified　by　the　second　stage　filtration.　After　that,　to　deal　with　the　concerned　problem,　an　improved　multi-objective　non-dominated　sorting　genetic　algorithm　III　(IMO-NSGA-III)　is　developed　to　find　the　Pareto-optimal　solutions.　Furthermore,　nine　random　instances　are　designed　to　show　that　the　proposed　IMO-NSGA-III　outperforms　other　three　state-of-the-art　algorithms　in　terms　of　convergence　and　diversity.　Finally,　three　case　studies　that　comes　from　an　automotive　parts　assembly　company　is　employed,　and　the　effectiveness　of　the　proposed　model　and　IMO-NSGA-III　algorithm　is　further　demonstrated.