In order to guarantee the timeliness of instant logistics, both soft time window and dynamic demands need to be considered. First of all, static distribution mode is established based on minimum timeout rate, travel distance and travel time of every order. By dividing the dynamic optimization stage into several dynamic service time windows with continuous time windows, the problem is transformed into a transient static VRPTW with correlation. Then, in the initial planning stage, which is static environment, an improved RTR algorithm is proposed based on the construction of the initial solution. In the second stage, which is dynamic environment, the order is firstly aggregated through the two attributes of the order's geographical location and time window to generate the aggregation path segment, and then the aggregation path segment is connected to form a new distribution path. Finally, a plug-in real-time path planning algorithm is proposed. An example is given to analyze and compare the solution of jsprit software package and the algorithm proposed in the paper. By comparing the results of static and dynamic environments, the economy of joint distribution and the advantages and disadvantages of the algorithm are analyzed, and the effectiveness of the real-time optimization algorithm proposed is verified.