The study evaluates the influence of presidential candidates’ personality traits on voting intentions, addressing a notable gap in political behavior research. Given Peru’s political institutions’ historical instability and fragmentation, understanding how voting operates in this context is relevant. The objective of the study is to examine the relationship between candidates’ personality traits and voter preferences using computer vision and deep learning techniques. The methodology involves using deep learning models to estimate the Big Five personality traits from video data of candidates during a simulated electoral process. The analysis utilizes two primary datasets: the First Impressions V2, containing labeled personality trait ratings to train the models, and a manually constructed database of Peruvian and international presidential candidates’ videos to determine personality traits from the trained models. Subsequently, surveys are conducted to track voting intentions in a simulated electoral process. The findings demonstrate significant correlations between specific personality traits, such as extraversion and agreeableness, and voting patterns, showing that personality traits can affect voter decisions, offering a more profound comprehension of voter decision-making processes and political dynamics to enhance communication strategies, evaluate political positions, and maintain party cohesion.