Context-aware systems (CAS) may face context con-flicts while collecting context data to support context-aware adaptation decisions. These systems need to implement mechanisms to resolve context conflicts in order to reduce the probability of making erroneous context-aware adaptation decisions. Quality of Context (QoC) indicators can be used to deal with this challenge, ensuring the correct behavior of context-aware systems. Indeed, context fusion layers of context management frameworks (CMF) should be able to detect and resolve context conflicts (internal and external) in order to maintain the consistency of contextsensitive decisions. Therefore, this paper presents a new approach to resolve internal and external context conflicts based on two QoC indicators: probability of correctness and trustworthiness. The proposed approach is embedded in the context fusion layer of our context management framework developed to support context-aware services and applications.