The Internet of Things (IoT) is a highly agile and complex environment managed via the Internet. The management of such an environment requires robust automated mechanisms, since manual curation becomes a prohibitively expensive and near-impossible task. Motivated by this observation, we present a mechanism to resolve conflicts among the services provided by IoT devices in such environments, by reformulating conflict resolution as a planning problem. Specifically, we propose to use planning with soft constraints (preferences) to resolve conflicts that arise when fulfilling multiple goals using varied services in an IoT environment. We build on previous work on creating a semantic policy framework for detecting conflicts within an IoT environment, and present a proof-of-concept implementation of our approach to demonstrate its promise in the IoT space.