Residence and work are two of the most important time and space behavior elements for?citizens.?To?a?great?extent,?commuting?pattern?reflects?spatial?structure?of?a?city.?Nowadays, the development of information and communication techniques provides new data sources and methodology for urban studies. This paper introduces former algorithms for calculating residence-and-work?anchor?points,?and?puts?forward?a?new?clustering?algorithm?for?internet LBS?data?based?on?DBSCAN.?A?case?with?the?data?produced?by?this?new?algorithm,?commuting patterns?of?eastern?Beijing?and?Beisanxian,?was?introduced?afterwards.?In?conclusion,?it’s found that the new algorithm for residence-and-work anchor points has satisfactory speed and accuracy,?and?is?suitable?for?the?application?of?LBS?data?in?urban?researches.