Understanding how individual personality traits relate to stress and abnormal behaviors is essential for improving animal welfare and management. However, few studies have quantitatively examined these relationships in giant pandas. This study aimed to investigate how personality traits predict stereotypic behavior and physiological stress responses in captive giant pandas under a simulated transport context. Stereotypies are abnormal repetitive behaviors thought to reflect stress coping, and personality is increasingly recognized as a factor shaping animal welfare. We compared non-stereotypic (n = 16) and stereotypic (n = 16) giant pandas using keeper-based ratings, crate-entry training, and urinary cortisol. Intraclass correlation and principal component analyses identified two personality components, activity and timidity. Non-stereotypic pandas were significantly less timid (p = 0.005), whereas activity did not differ (p = 0.355). In crate-entry training, non-stereotypic pandas tended to enter faster (48.5 vs. 139.5 s; p = 0.074), and timidity was strongly correlated with latency (ρ = 0.74, p < 0.01). Censoring-aware analyses treating 300 s as right-censoring further supported this pattern (stratified Cox (Hazard Ratio) HR = 3.09, 95% CI 0.93–10.25; likelihood-ratio p = 0.052; log-normal accelerated failure-time model time ratio = 0.36, p = 0.054). Urinary cortisol showed no between-group difference after adjustment (geometric mean ratio = 0.93, 95% CI 0.57–1.49; p = 0.75), with consistent sensitivity analyses. These findings indicate that stereotypic pandas are more timid, and higher timidity predicts slower adjustment to a stress context, highlighting the value of personality assessment in management to mitigate stereotypies and improve welfare, while inconsistent cortisol results warrant further study.

Environmental enrichment is commonly used in captive animal management, but the behavioural effects of specific enrichment devices are not always evaluated systematically. We conducted two experiments to evaluate three foraging-based feeding devices used with adult male giant pandas during summer indoor housing. In a repeated-presentation experiment with four pandas, responses to the devices were strongly individual-specific: three pandas interacted most with Device B and one with Device A, and interaction duration declined across repeated presentations in a non-uniform manner. In a complementary exploratory single-case (A–B) comparison with two pandas, feeding devices were associated with a marked within-individual reduction in abnormal behaviour (90.2% and 75.1%; Cliff’s delta = −0.83 in both) and a shift in behavioural events from locomotion toward exploration. These findings support evaluating feeding enrichment at the individual level in captive giant panda management.