In　this　paper,　based　on　spatial　autoregression　and　partial　functional　linear　regression,　we　introduce　a　new　varying-coefficient　partial　functional　spatial　autoregressive　model.　The　functional　principal　component　analysis　and　B-spline　are　adopted　to　approximate　the　slope　function　and　varying-coefficient　functions　respectively.　Then,　the　instrumental　variable　method　gives　final　estimators.　Under　some　regular　conditions,　we　further　study　the　asymptotic　normality　of　the　parameter　and　the　convergence　rates　of　slope　function　and　coefficient　functions.　Lastly,　the　finite　sample　performance　of　the　proposed　methodology　is　evaluated　by　simulation　studies　and　a　practical　data　example.