In order to solve the problem of the truck-drone coordinated delivery path planning, which involves the simultaneously pickup and delivery while reducing carbon emissions, the objective is set to minimize the total transportation cost with constraints such as the balance of entries and exits of trucks and drones, the unique access and traversal, load limit, endurance limit, and the absence of self-loop, and a path planning model for truck-drone joint pickup and delivery considering carbon emission is established. The K-means-improved ant colony optimization two-stage algorithm is designed to obtain the approximate optimal solution of the model. The K-means algorithm is used to cluster customer points, clustering center points serve as truck stop locations, and the customer points within a cluster to which a clustering center belongs serve as drone stop locations. On the basis of the original transition probability formula of the ant colony optimization algorithm，a new probability selection factor is introduced，which increases the probability of paths of customer points with smaller net pickup quantities being selected. Simulation results show that compared to solo truck transport, the total transportation time of the truck-drone transport mode is shorten by 5.04% on average, the total transportation cost is reduced by 15.56% on average, and the carbon emission is reduced by 32.50% on average. In most scenarios, the proposed model can solve the problem quickly； especially in the scenario where customer points are evenly distributed, the effect is even better.