Time　synchronization　of　clocks　in　the　sensor　nodes　for　wireless　sensor　networks　(WSNs)　is　a　fundamental　technology　for　most　mission-critical　applications.　Most　of　past　research　in　time　synchronization　for　WSNs,　however,　has　only　focused　on　achieving　some　of　the　goals　at　a　time,　such　as　accuracy,　energy　consumption,　completion　time,　etc.,　making　these　solutions　less　capable　of　adapting　to　different　application　requirements.　In　this　paper,　we　propose　a　new　time　synchronization　algorithm　named　MBATS　(mobile　beacon-based　adaptive　time　synchronization)　in　which　a　mobile　beacon　is　employed　to　move　or　fly　over　the　sensor　deployment　area　to　complete　time　synchronization.　Moreover,　MBATS　is　designed　so　that　the　number　of　sensor　nodes　that　are　synchronized　by　one　instance　of　time　synchronization　from　the　mobile　beacon　could　vary　dynamically　to　meet　application　requirements　on　accuracy,　completion　time　and　energy　consumption,　making　the　proposed　MBATS　algorithm　highly　adaptable　to　different　application　requirements.　In　addition　to　showing　the　advantage　of　the　proposed　MBATS　algorithm　on　the　adaptability　of　time　synchronization　as　well　as　on　some　of　the　main　metrics　of　synchronization　over　comparable　schemes　for　WSNs,　we　also　present　the　results　of　our　study　on　comparing　the　performance　of　letting　the　mobile　beacon　traverse　along　a　designing　path　versus　follow　a　random　path.　Such　a　study　is　important　since　it　would　allow　us　to　learn　the　performance　gains　that　we　can　expect　to　achieve　with　extra　control　effort　spent　on　designing　the　path　over　the　effortless　random　path　strategy.　Such　study　could　provide　us　with　some　clues　on　how　to　choose　a　suitable　time　synchronization　strategy　to　better　meet　application　requirements,　which　may　not　necessarily　be　the　designed　path　strategy　due　to　the　tradeoff　between　cost　and　performance　gains.