The geographic information data of mega city has the characteristics of massive volume, rapid accumulation, heterogeneous and being frequent used. In the era of big data and Web2.0, traditional storage and management of spatial data and sharing service mode lacks of data integration and mining, as well as passively serves customers without differences. It is difficult to meet the increasing requirements of richer spatial data content, quicker response and higher quality. And it is difficult to adapt to individualized demand of spatial information service. In this paper, the key technologies of personalized services and cloud spatial data centre construction have been researched, personalized situational data application mode is proposed, a method of online heterogeneous space data integration has been described, individual spatial data and traditional spatial data service integration is realized based on NoSQL. Taking Beijing as an example, a mega cities space cloud data centre based on private cloud has been built, which is reliable, efficient, expandable and has high participation degree and provides personalized spatial services.