Background: Stress is a major risk factor for the impairment of psychological well-being. The present study aimed to evaluate the empirical evidence of the Transactional Stress Model proposed by Lazarus and Folkman in patients with psychosomatic health conditions. Methods: A structural equation model was applied in two separate subsamples of inpatients from the Department of Psychosomatic Medicine (total n = 2,216) for consecutive model building (sample 1, n = 1,129) and confirmatory analyses (sample 2, n = 1,087) using self-reported health status information about perceived stress, personal resources, coping mechanisms, stress response, and psychological well-being. Results: The initial model was created to reflect the theoretical assumptions by Lazarus and Folkman about their transactional stress concept. This model was modified until a sufficient model fit was reached (sample 1: CFI = 0.904, TLI = 0.898, RMSEA = 0.072 [0.071–0.074], SRMR = 0.061). The modified model was confirmed in a second sample (sample 2: CFI = 0.932, TLI = 0.928, RMSEA = 0.066 [0.065–0.068], SRMR = 0.052). Perceived external stressors and personal resources explained 91% of the variance of the stress response, which was closely related to symptoms of depression (63% variance explained). The attenuating effect of resources on stress response was higher (standardized β = -0.73, p < 0.001) than the impact of perceived stressors on stress response (standardized β = 0.34, p < 0.001). Conclusion: The empirical data largely confirmed the theoretical assumption of the Transactional Stress Model, which was first presented by Lazarus and Folkman, in patients with a wide range of psychosomatic conditions. However, data analyses were solely based on self-reported health status. Thus, proposed inner psychological mechanisms such as the appraisal process could not be included in this empirical validation. The operationalization and understanding of coping processes should be further improved.