Image　interpolation　is　a　key　aspect　of　digital　image　processing.　This　paper　presents　a　novel　interpolation　method　based　on　reproducing　kernel　filter(RKF)　and　global　texture　orientation.　With　the　kernel　function　in　H1(K)　space,　it　can　be　shown　that　the　RKF　is　constructed　and　a　one-dimensional　interpolation　function　is　derived.　According　to　H1(R)　norm,　the　RKF　has　the　advantage　of　offering　good　approximation　of　the　function　values　and　its　first　derivatives　at　the　knots.　A　two-dimensional(2-D)　Gabor　filter　is　used　to　compute　local　features　which　characterizes　directional　behavior　of　the　texture　and　decide　interpolation　direction.　Simulation　results　of　numerous　test　image　show　that　the　proposed　technique　is　subjectively　and　objectively　far　better　than　other　existed　method.　©　2006　IEEE.