Ever　since　the　concept　of　swarm　intelligence　was　brought　out,　a　variety　of　control　algorithms　for　swarm　robotics　has　been　put　forward,　and　many　of　these　algorithms　are　stable　enough　and　efficient.　Most　of　the　researches　only　take　an　invariable　controller　which　functions　through　the　whole　stage　into　consideration,　the　situation　in　which　controller　changes　over　time　is　rarely　taken　into　account.　However,　there　are　limitations　for　invariable　controller　dominated　algorithms　in　practical　situation,　which　makes　them　unable　to　meet　changing　environment.　On　the　contrary,　variable　controller　is　more　flexible　and　can　be　able　to　adapt　to　complex　environment.　Considering　such　advantages,　a　time-varying　algorithm　for　swarm　robotics　is　presented　in　this　paper.　The　algorithm　takes　time　as　one　of　the　independent　variables　so　that　the　controller　is　no　longer　fixed　through　the　time,　but　can　be　changed　over　time,　which　brings　more　choices　for　the　swarm　robot　system.　In　this　paper,　some　relevant　simulations　are　designed　to　test　the　algorithm.　Different　control　strategies　are　applied　on　the　same　flock　during　the　time,　and　a　more　complex,　flexible　and　practical　control　effect　is　acquired　successfully.