With　the　rapid　development　and　popularization　of　smartphones　and　cloud　computing,　computer-generated　screen　content　images　(SCIs)　are　important　mediums　that　represent　enormous　Internet　information,　having　penetrated　into　our　daily　lives.　There　are　many　natural　scene　image　quality　assessment　(QA)　methods　proposed　in　the　past　decades,　but　these　methods　are　improper　for　SCIs　since　SCIs　include　three　complex　contents,　i.e.,　texts,　graphics,　and　illustrations,　where　distortion　results　in　varying　degrees　of　degradation.　For　the　QA　of　SCIs,　this　chapter　first　presents　the　full-reference　QA　methods　based　on　structural　similarity　to　estimate　structural　changes　and　different　statistical　properties　of　regions.　Second,　for　the　sake　of　tackling　the　issue　of　unsatisfactory　prediction　monotonicity,　it　introduces　the　reduced-reference　QA　method　based　on　the　fusion　of　macroscopic　and　microscopic　features.　Third,　to　deal　with　the　difficulties　of　monotonous　color　and　simple　shape　in　SCIs,　it　shows　the　no-reference　QA　methods　based　on　adaptive　multi-scale　weighting　and　big　data　learning.　In　the　end,　the　future　research　trends　of　SCI　QA　are　discussed,　and　the　necessity　of　constructing　accurate　and　efficient　objective　QA　models　of　SCIs　is　indicated.　©　2022,　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.