Fear of threatening contexts often generalizes to similar safe contexts, but few studies have investigated how contextual information influences cue generalization. In this study, we explored whether fear responses to cues would generalize more broadly in a threatening compared to a safe context. Forty-seven participants underwent a differential cue-in-context conditioning protocol followed by a generalization test, while we recorded psychophysiological and subjective responses. Two faces appeared on a computer screen in two contexts. One face (CS+) in the threat context (CTX+) was followed by a female scream 80% of the time, while another face (CS-) was not reinforced. No faces were reinforced in the safe context (CTX-). In the generalization test, the CSs and four morphs varying in similarity with the CS+ were presented in both contexts. During acquisition, conditioned responses to the cues were registered for all measures and the differential responding between CS+ and CS- was higher in CTX+ for US-expectancy ratings and skin conductance responses, but the affective ratings and steady-state visual evoked potentials were not context-sensitive. During test, adaptive generalized responses were evident for all measures. Despite increased US-expectancy ratings in CTX+, participants exhibited similar cue generalization in both contexts, suggesting that threatening contexts do not influence cue generalization.