With the emergence of connected and automated vehicles (CAVs), the mixed traffic environment where conventional human-driven vehicles (HVs) share the road space with CAVs is impending. Targeted management of such mixed traffic flow at intersections is of practical significance, inspiring the theoretical exploration of CAVs impacts on intersection capacity in this study. Penetration rate and platooning behavior of CAVs are the major concern. Not only the three core platooning parameters (intra-platoon gap, inter-platoon gap and platoon size) are investigated, but also the individual willingness of CAVs to form platoons is innovatively considered. Intersection capacity is modeled as an objective function of a linear program problem that maximizes the sum of throughput of each stream crossing the intersection under collision-free constraints. These constraints are determined by the conflicting relationship of vehicles on different streams and the vehicle’s average occupation time of conflict points. The average occupation time is defined by adopting the service time concept in queueing theory, and specified by applying the Markov chain theory considering the spatial distribution of heterogeneous vehicle headways in mixed traffic. This study provides a practical procedure to analyze the influence of CAVs on mixed traffic at intersections, without the need to determine specific intersection management methods. Corollaries drawn are basically consistent with expectations, except that high CAV penetration does not necessarily correspond to high intersection capacity. In order for the influence of penetration to be positive, the platooning willingness must be controlled within the corresponding range, suggesting that CAV penetration rate and platooning willingness should be considered jointly when updating intersection management for mixed CAVs–HVs traffic flow. Furthermore, numerical analyses are carried out to corroborate the analytical conclusions. •Analytical formulation for intersection capacity by integrating linear program and queueing theory is developed for mixed traffic with CAVs.•CAV penetration rate and related heterogeneous vehicle headways in mixed traffic are considered.•CAVs platooning parameters including willingness contribute to the intersection capacity.•Analytical analysis is verified by Monte Carlo simulations.