The　development　of　cloud　computing　makes　it　possible　to　dynamically　adjust　resources　for　software　services　on　demand.　However,　different　software　services　have　different　QoS　requirements,　and　their　external　environment　is　changing　all　the　time.　Thus,　self-adaptive　ability　is　needed　for　software　services　because　engineers　manage　is　difficult　in　the　cloud　resources.　This　paper　proposes　a　self-adaptive　resource　management　framework　for　software　services　in　Cloud.　First,　for　a　given　software　services,　its　iterative　QoS　model　is　trained　on　historical　data,　which　is　capable　to　predict　a　QoS　value　of　one　management　operation　by　using　the　information　on　current　running　workload,　allocated　resources,　real　QoS　value　and　an　operation　of　resource　allocation.　Then,　we　employ　PSO-based　runtime　decision　algorithm　together　with　the　predicted　QoS　value　to　determine　future　resource　allocation　operations.　Last,　a　feedback　control　loop　is　introduced　to　gradually　allocate　resource　through　feedback　and　iteration.　The　loops　iterate　until　the　PSO-based　algorithm　suggests　no　further　improvement　over　the　current　resource　allocation.　Such　self-adaptive　management　programs　can　adapt　to　the　change　in　the　outer　environments　and　allocate　cloud　resource　effectively.　©　2019　IEEE.