In order to resolve the problem that the volume of APP-based car-hailing industry increases and the trip counting results are susceptible to errors due to environmental influences, a novel method of APP-based car-hailing trip counting detection was proposed. The conversion mechanism between global navigation satellite system(GNSS) coordinate system and other coordinate systems was analyzed, and a network car counting device was designed by using loose coupling model and Kalman filter processing for combined navigation. The test results show that the distance counting device has a distance counting error of 0.42% in the straight road section and a distance counting error of 0.56% in the circular road section, and the distance counting accuracy meets the maximum error range required by the Verification Regulation of Taximeters(JJG 517—2016), which provides a new method for solving the distance counting error problem of APP-based car-hailing.