The　supercritical　carbon　dioxide　(S-CO2)　Brayton　cycle　is　a　potential　technology　for　recovering　waste　heat　from　compressed　natural　gas　(CNG)　engines.　To　choose　suitable　cycle　configurations　is　important　considering　the　complex　operating　characteristics　of　CNG　engines.　The　traditional　evaluation　models　of　the　S-CO2　cycles　configuration　are　relatively　one-sided　without　the　comprehensive　considerations　of　evaluation　metrics.　On　the　basis　of　the　analytic　hierarchy　process　and　information　entropy　theory,　this　paper　proposes　a　multi-dimensional　comprehensive　evaluation　method　for　S-CO2　cycles　from　three　perspectives:　thermodynamic,　economic,　and　environmental　performance.　A　thermoeconomic　performance　optimization　is　conducted　using　a　non-dominated　genetic　algorithm　II.　Unlike　most　of　the　existing　researches　that　use　a　single　stable　operating　condition,　this　paper　considers　the　full　operating　conditions　of　engines.　Results　show　that　the　recompression　cycle　has　the　best　overall　performance　among　the　investigated　cycles　with　an　exergy　efficiency　and　electricity　production　cost　of　67.20　%　and　0.45　$/kWh,　respectively.　The　efficiency　improvement　of　the　CNG　engine　coupled　with　the　WHR　system　can　reach　up　to　6.31　%.　This　study　can　provide　a　new　reference　for　the　comprehensive　multidimensional　evaluation　and　performance　characteristics　and　optimal　performance　acquisition　under　full　engine　operating　conditions　of　the　S-CO2　cycle.