Abstract: In order to achieve personalized and intelligentialized POI recommendations for Web mapsand adapt services to the users’different requirements,we discuss the context factors of POI person-alized recommendations and analyze multidimensional contexts which can influence the personal needsof users.On the basis of research concerning the context-based,methods and process for POI recom-mendations,an improved algorithm for TOPSIS personalized recommendation,in which the utility ofthe service is taken into consideration,is proposed.In this study,discussions focus on the treatmentstrategy for multidimensional contexts(such as geographic space and time)in a recommendation algo-rithm.An experiment demonstrates that the algorithm is feasibile and valid when generating recom-mendation results.