Predicting　and　analyzing　the　availability　of　green　space　resources　for　the　elderly　are　crucial　for　improving　their　quality　of　life　and　addressing　the　challenges　of　an　aging　society.　Taking　Xining　City,　the　area　with　the　largest　elderly　population　in　Qinghai　Province,　as　the　study　area,　this　study　identifies　and　predicts　the　spatial　distribution　of　older　people　based　on　mobile　phone　location　data　with　age　identification　information.　The　Gaussian-based　two-step　floating　catchment　area　(G2SFCA)　method　was　then　employed　to　study　and　predict　the　accessibility　of　parks　and　green　spaces　for　the　elderly.　The　following　results　were　observed:　(1)　The　elderly　population　increasing　rate　exhibited　a　circular　distribution,　displaying　a　low　rate　in　the　central　city　and　outer　suburbs　and　a　high　rate　in　the　inner　suburbs.　(2)　The　overall　spatial　distribution　pattern　of　accessibility　of　parks　and　green　spaces　for　the　elderly　did　not　change　significantly　over　the　10-year　forecast　period,　but　the　general　accessibility　level　declined.　Under　the　condition　of　walking　for　15　min,　the　population　covered　by　relatively　high　and　high　grades　of　accessibility　decreased　from　17.58%　to　6.70%.　Moreover,　under　the　condition　of　public　transportation　for　30　min,　the　population　covered　by　relatively　high　and　high　grades　of　accessibility　decreased　from　26.41%　to　9.28%.　(3)　It　was　found　that　the　relative　variability　of　accessibility　of　parks　and　green　spaces　for　the　elderly　is　significant　from　2018　to　2028,　with　approximately　87%　of　the　parks　and　green　spaces　experiencing　a　reduction　of　＞70%　in　accessibility　under　both　walking　and　public　transportation　conditions　for　30　min.　This　study　provides　valuable　insights　for　future　urban　park　and　green　space　planning,　particularly　in　response　to　the　needs　of　an　aging　population.　©　2023　Authors.　All　rights　reserved.