The opioid crisis in the United States has reached epidemic levels, with the number of opioid-related overdose deaths now surpassing car accident fatalities. The severity of this crisis makes identifying the key risk factors associated with opioid overdose behaviors crucial. This research demonstrates the importance of investigating personality traits as a risk factor for opioid overdose mortality. We use the five-factor personality model in conjunction with advanced text mining to infer personality traits from unstructured social media data at the county level. We then employ econometric modeling to estimate the relationships between the inferred personality traits and fatal opioid overdose. Based on the results, conscientiousness and extroversion were found to be negatively associated with fatal opioid overdose, while neuroticism was positively associated with it. Furthermore, a counterintuitive finding suggested that agreeableness had a positive effect. Our results yield several practical implications that can complement existing opioid surveillance practices.