The　quality　of　traffic　data　collected　by　traffic　detectors　for　open　highways　is　easily　affected　by　bad　weather　and　road　construction,　so　the　quality　of　data　needs　to　be　evaluated.　Considering　that　macroscopic　evaluation　method　requires　a　long　period,　leading　to　ineffective,　abnormal　data　quality　detection,　this　paper　proposes　a　method　combining　online　evaluation　with　offline　evaluation　method　of　traffic　data　quality.　The　indices　of　online　evaluation　method　include　threshold　of　volume,　speed,　occupancy,　and　the　consistent　relationships　among　the　traffic　parameters　that　are　proposed　for　practical　engineering　applications;　while　offline　evaluation　indices　consist　of　volume　daily　variation,　absence　rate,　and　error　rate.　This　paper　designs　a　repair　process　considering　the　characteristics　of　temporal　correlation,　spatial　correlation,　and　historical　correlation.　The　experimental　results　indicate　that　the　proposed　evaluation　method　can　be　carried　out　and　has　a　practical　significance　for　the　maintenance　of　traffic　detectors.　©　2018　American　Society　of　Civil　Engineers.