Using the measured data to statistically analyze traffic parameters and then construct traffic load models is a popular approach to the estimation of the actual bridge load effect of random motorcade on bridge． However， the rationality of such an approach has not been theoretically verified till now． In this paper，based on the vehicle information collected by the weigh-in-motion systems ( WIM) on 28 lanes of six highway bridges，the 1-D random field of vehicle type of traffic flow on each lane is established． Then，the numerical characteristics of the established random fields，such as the ensemble means，the collection of correlation functions，the spatial means and the spatial correlation functions，are calculated according to the theory of random process． Analytical results show that the 1-D random fields of vehicle type of traffic flow，which are established through the survey on adequate vehicle samples on actual highways，are of stationarity and ergodicity，and the statistical information of the samples can represent the total samples． Besides，the established random fields can be used to further investigate the random vehicle loads and can provide a theoretical basis for modeling the random vehicle load on highway bridges．