Job　mobility　is　common　in　modern　society,　especially　in　the　industry　of　information　and　communication　technology.　Job　mobility　prediction　is　valuable　both　for　employees　and　employers.　For　the　sake　of　lacking　appropriate　and　sufficient　records　of　job　mobility,　the　traditional　methods　meet　a　significant　challenge　in　job　mobility　prediction.　Fortunately,　the　emerging　professional　social　network　provides　a　large　amount　of　users＇　career　histories,　which　can　alleviate　this　problem.　In　this　paper,　we　collect　relevant　data　from　LinkedIn,　and　analyze　the　temporal　and　spatial　characteristics　to　model　the　job　mobility　pattern.　We　propose　an　approach　to　predict　the　company　size　and　position　for　the　next　job　by　using　various　features,　such　as　the　position,　duration,　and　size　of　previous　companies,　education　degree,　etc.　The　experimental　results　verify　the　proposed　approach　with　the　accuracy　up　to　72%　and　74%　in　terms　of　next　company　size　prediction　and　next　position　prediction　respectively.　©　2016　IEEE.