Software-Defined WSNs (SDWSNs) have attracted considerable attention as they can provide more flexible network management compared with traditional WSNs. However, they also bring a new security issue, i.e., a sensor node can be easily compromised and behave maliciously to perform arbitrary action, e.g., dropping received messages, that degrades the availability of SDWSNs without being detected. To address the security issue, we propose a Hierarchical Trust Management scheme, named HTM. In HTM, a reputation-based mechanism is designed and utilized for detecting sensor nodes’ malicious behavior, such as black-hole attack (dropping all received packets), selective forwarding attack (dropping partially received packets and forwarding the rest), and Denial-of-Service (DoS) attack (sending abundant but useless packets continuously). At each level of the hierarchical system, the trustworthy of each node is evaluated and the malicious behavior is detected. Through extensive simulation, we demonstrate that the HTM scheme is capable to detect malicious nodes that perform the aforementioned attacks.