Mood congruity and affective priming have been used to study the effects of affective phenomena on perception. Manipulation of mood is appropriate for investigations of long-term effects while affective priming is limited to short intervals (approximately 300 ms) between a prime and target. However, studying the influence of real-world rapidly changing emotional episodes on perception may fall between the cracks of these methods. This may be caused, inter alia, because emotional episodes are distinguished from mood experiences on one hand, but often last for longer than roughly 300 ms on the other. Thus, it is unclear what experimental approach should be taken to investigate congruency effects triggered by emotional episodes. The present study used a new variation of the evaluation decision task (EDT) combined with a script-driven imagery procedure to investigate a possible congruency relationship between the evaluator's emotional experience at a given time and observable emotional markers of others. We used 180 9-word script-driven imageries as varying valence primes (negative, positive, neutral) and asked participants to imagine themselves in the situation described in the scripts. At the last stage of a trial, all participants were asked to evaluate the mood—positive or negative—of a target face of a child in a photograph. We manipulated the reading interval (4000 ms and 1350 ms) and the subsequent blank interval (300 ms, 5000 ms, and unlimited) until target onset. Prime and target valence were congruent or incongruent. Significant congruency effects were found for both short and long reading intervals and blank intervals. However, in longer blank intervals only the interference effect reached significance. Furthermore, the interference effect was found to be significant mainly in trials beginning with a negative script.