Through the lens of preclinical research on substance use disorders (SUD), I propose a reflection aimed at reevaluating animal models in neuroscience, with a focus on ecological relevance. While rodent models have provided valuable insights into the neurobiology of SUD, the field currently faces a validation crisis, with findings often failing to translate into effective human treatments. Originally designed to address the lack of reproducibility in animal studies, the current global gold standard of rigorous standardization has led to increasingly controlled environments. This growing disconnection between laboratory settings and real-world scenarios exacerbates the validation crisis. Rodent models have also revealed various environmental influences on drug use and its neural mechanisms, highlighting parallels with human behavior and underscoring the importance of ecological relevance in behavioral research. Drawing inspiration from inquiries in ethology and evolutionary biology, I advocate for incorporating greater environmental complexity into animal models. In line with this idea, the neuroethological approach involves studying spontaneous behaviors in semi-natural habitats while utilizing advanced technologies to monitor neural activity. Although this framework offers new insights into human neuroscience, it does not adequately capture the complex human conditions that lead to neuropsychiatric diseases. Therefore, preclinical research should prioritize understanding the environmental factors that shape human behavior and neural architecture, integrating these insights into animal models. By emphasizing ecological relevance, we can achieve deeper insights into neuropsychiatric disorders and develop more effective treatment strategies. This approach highlights significant benefits for both scientific inquiry and ethical considerations.