As eating behavior is important to health, this cross-sectional study was conducted to analyze the factors influencing the eating behavior related to overweight and obesity of Chinese residents aged 18~60 based on the Ecological Model of Health Behavior. The short-form of the Eating Behavior Scale (EBS-SF) was applied to evaluate eating behavior. The multivariable linear stepwise regression analysis was used to identify and analyze the influence factors, and the receiver operating characteristic curves analysis to validate the predictive capability of the EBS-SF score in differentiating overweight and obesity. A total of 8623 participants were enrolled. In the personal characteristics, male (β = −0.03), older [36–45 years (β = −0.06) or 46–60 years (β = −0.07)], higher scores of Agreeableness (β = −0.04), Conscientiousness (β = −0.14) or Openness (β = −0.03) contributed to healthy eating behavior. In the individual behaviors, those who smoked (β = 0.04), drank alcohol (β = 0.05), exercised frequently (β = 0.07), had higher PHQ-9 scores (β = 0.29) may have improper eating habits. As for the interpersonal networks, the residents who were married (β = −0.04) behaved well when eating, while those who had offspring or siblings tended to have unhealthy eating behavior. At the community level, living in Western China (β = −0.03), having a monthly household income of 6001–9000 yuan per capita (β = −0.04), having no debt (β = −0.02), being retired (β = −0.03), or having lower PSSS scores (β = −0.03) led to lower EBS-SF scores. And the EBS-SF score demonstrated a moderate-high accuracy in predicting overweight and obesity.