Abstract: Probe vehicle data are increasingly becoming the primary source of traffic data. In current practice, traffic volumes and speeds are collected from inductive loop or similar devices. As probe vehicle data become more widespread, it is imperative that methods are developed so that traffic state estimators like speed, density and flow can be derived from probe vehicle data as well. In this paper, a methodology to estimate traffic flow on a freeway based on probe vehicle trajectory data combined with traffic shockwave theory is proposed. In essence, probe vehicle trajectory can indicate the free-flowing and congested regimes. By using LWR kinematic wave model, a shockwave can be identified that separates both regimes. From the formation of the shockwave, flows for each regime are estimated. To identify the shockwave, k-means clustering is applied to the data. When applied to simulated data, the error of the estimated flow during free-flow ranges from -9% to 1% with an average of -5%. The estimated flow during congestion has an error of 0%. Based on the results, this paper shows that the proposed method can predict traffic flow with a reasonable accuracy under congested and free-flow conditions.