A　robotic　neuro-control　that　connects　the　neural　network　and　PI　control　to　construct　a　hybrid　scheme　of　robotic　inverse-model　learning　control　has　been　proposed　for　trajectory　tracking　control　free-floating　space　robot　with　rigid　manipulators　and　ridid-flexible　manipulators,　and　the　working　principles　of　control　are　demonstrated.　Finally,　a　simulation　of　control　scheme　for　a　free-floating　space　robot　with　two-manipulator　was　made.　The　simulation　results　have　shown　that　the　proposed　inverse-model　learning　neuro-control　possesses　great　speed　of　convergence　and　higher　tracking　precision,　and　can　provide　an　effective　control　for　free-floating　space　robot.