With　the　explosive　growth　of　emerging　applications　in　wireless　networks,　high　transmission　rate,　low　latency　and　secure　data　and　resource　sharing　have　become　urgent　requirements.　Mobile　edge　computing　(MEC)　technology　can　provides　low　latency　computing　services　and　high　speed　content　cache　to　obtain　a　higher　service　quality.　In　combined　with　blockchain　technology,　which　are　capable　of　establishing　trust　in　an　untrustworthy　environment,　the　security　function　of　MEC　can　be　enhanced　and　satisfies　the　new　application　needs.　In　this　paper,　the　data　sharing　application　in　mobile　edge　blockchain　is　considered.　The　BS　is　responsible　of　forwarding　data　package　from　the　source　to　the　target　UE.　Meanwhile,　the　MEC　server　deployed　at　the　BS　generates　a　transaction　recording　this　sharing　event　and　mine　the　corresponding　block　to　ensure　security.　The　profit　of　the　BS　related　to　the　wireless　data　transmission　cost　of　frequency　resource　and　power,　the　mining　reward　and　the　edge　computing　cost　of　the　transaction　and　the　competitiveness　of　the　MEC　server　in　the　blockchain　system　is　defined　as　the　network　utility.　The　data　sharing　task　scheduling　problem　at　the　BS　is　modeled　as　a　Markov　decision　process　(MDP),　and　policy　gradient　based　task　scheduling　(PG-TS)　algorithm　is　proposed　and　evaluated.　Simulation　results　show　that　the　performance　of　the　proposed　algorithm　is　better　than　greedy,　random,　and　first　in　first　out　(FIFO)　algorithm.　©　2020　IEEE.