Considerable　errors　exist　in　the　road　speed　data　obtained　from　various　detectors　and　directly　influence　the　accuracy.　Various　traffic　models　were　established　for　estimation.　12　groups　of　field　experiments　on　different　classes＇　roads　by　using　a　test　vehicle　with　global　positioning　system　(GPS)　modules　were　implemented.　Through　combining　automatically　recording　and　manually　recording,　the　accuracy　discrepancy　of　travel　speed　data　collected　by　floating　car　system,　microwave　detectors,　and　loop　detectors　was　attempted　to　verify.　The　results　analysis　shows　that　the　average　error　rate　of　travel　speed　of　floating　cars　at　road　section　on　urban　expressways　and　arteries　are　13.6%　and　27.8%,　and　the　acquisition　data　and　the　actual　speed　have　the　same　distribution.　Microwave　detector　on　expressways　and　loop　detectors　on　arteries　are　30%　and　56%.　Even　though　the　section　velocities　of　the　microwave　detector　have　some　errors,　which　could　more　or　less　describe　the　basic　driving　conditions.　Thus,　results　provide　some　evidences　for　selecting　appropriate　traffic　model　decided　by　accuracy.　Moreover,　it　provides　evidences　for　the　application　of　different　dimensions　information　in　multi-source　data　fusion　for　higher　accuracy.