This paper studies the problem of pickup pharmacy selection and rider routing for O2O (online to offline) drug delivery. Based on the actual situation where a single pharmacy holds limited types of drugs and each customer order requires complete delivery, a collaborative optimization model of pickup pharmacy selection and rider routing for O2O drug delivery is formulated, where the total cost of the platform is minimized. And according to the complexity characteristics of the model, a nested algorithm of artificial immunity and ant colony optimization is designed. The analysis results of multiple sets of different scales and real-life scenarios show that the collaborative optimization of pickup pharmacy selection and rider routing for O2O drug delivery can effectively improve the utilization of distribution resources and decrease the total cost of the platform. The change in the number of drug types has little impact on the number of riders, but has a certain positive impact on the cost of distribution routes and the total cost of the platform. An increase in sufficient inventory of drugs in chain pharmacies can reduce the total cost of platform distribution and order splitting rate. The research conclusions can provide a useful reference for platforms to formulate O2O drug order pickup and delivery schemes.