BACKGROUND AND OBJECTIVES We argue that the stimuli used in traditional fear conditioning paradigms are too simple to model the learning and unlearning of complex fear memories. We therefore developed and tested an adapted fear conditioning paradigm, specifically designed for the study of complex associative memories. Second, we explored whether manipulating the meaning and complexity of the CS-UCS association strengthened the learned fear association. METHODS In a two-day differential fear conditioning study, participants were randomly assigned to two experimental conditions. All participants were subjected to the same CSs (i.e., pictures) and UCS (i.e., 3 s film clip) during fear conditioning. However, in one of the conditions (negative-relevant context), the reinforced CS and UCS were meaningfully connected to each other by a 12 min aversive film clip presented prior to fear acquisition. Participants in the other condition (neutral context) were not able to make such meaningful connection between these stimuli, as they viewed a neutral film clip. RESULTS Fear learning and unlearning were observed on fear-potentiated startle data and distress ratings within the adapted paradigm. Moreover, several group differences on these measures indicated increased UCS valence and enhanced associative memory strength in the negative-relevant context condition compared to the neutral context condition. LIMITATIONS Due to technical equipment failure, skin conductance data could not be interpreted. CONCLUSIONS The fear conditioning paradigm as presented in the negative-relevant context condition holds considerable promise for the study of complex associative fear memories and therapeutic interventions for such memories.