We focus on the topology control algorithm in Software-Defined Sensor Networks (SDSNs) and propose a robot-assisted sink nodes relocation algorithm based on the Markov chain. Firstly, by using the software-defined architecture, sensor nodes and network topology can be deployed and updated by the controller which provides global network information and centralized management in real time. Then robot-assisted sink nodes are deployed into SDSN to extend the network lifetime. Unlike motionless nodes or random mobile nodes which may bring about unforeseeable network topological variation, robot-assisted sink node can be migrated to specified location by the controller in SDSN. Furthermore, we propose a Markov chain based Robot-assisted Weighted Relocation (MRWR) algorithm to redeploy the sink nodes on the basis of node status and weight. Simulation results and analysis indicate that the proposed algorithm extends network lifetime with lower average power consumption and decreases the average delay from sensor nodes to sink node as compared with that of static network.