As a part of the smart urban construction, automated driving is introduced to improve the utilization efficiency of cars and roads, which not only reduces the incidence of traffic accidents, but also improves the environment quality. With the development of the smart urban, it is predictable that, in the city of the future, the service of package pickup and delivery or takeout will be supported mainly by automated vehicles. However, the existing works mainly focus on the variants of the Vehicle Routing Problem (VRP), in which they either take no account of service time of automated vehicle for customers when the automated vehicle arrives at the locations of customers or ignore the impact of rewards gained from customers on path planning of the automated vehicles. In this paper, we also extend a variant of VRP where an automated vehicle is used to package delivery or distribution of food in the smart urban environment, which is called the Delivery Reward Maximization (DRM) problem. The problem aims at designing a route of the automated vehicle while considering the service time for customers before their deadlines and the impact of rewards of the automated vehicle on path planning. We first prove that the DRM problem is NP-hard. Then we study two special cases of the DRM problem, which are called Linear DRM (LDRM) problem and Two-dimensional DRM (TDRM) problem, respectively. In the LDRM and TDRM problems, all customers have the same visiting deadlines and are deployed on the one-dimensional line and two-dimensional plane, respectively. Then we prove that the LDRM and TDRM problems are also NP-hard and propose a constant approximation algorithm for each of them. Afterward, we propose a greedy algorithm to solve the DRM problem, and give the analysis by counterexample.