OBJECTIVES: To validate a model for quantifying the COPD disease progression against both the data used to generate the model (internal validation) and clinical trial data not used in the model’s development (external validation).  METHODS: A model representing causal relationships between central disease attributes (lung function, exacerbations, symptoms and exercise capacity) and final outcomes (survival, quality of life, cost) was developed based on the Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) study dataset. Model predicted annual outcomes were compared to the corresponding annual observed data from ECLIPSE (n=2,164) and TOwards a Revolution in COPD Health (TORCH) (n=6,108) trials based on fitting the model baseline parameters to reflect each specific study population.  RESULTS: The model accurately predicted the ECLIPSE outcomes in at least two of the three annual time points within the 95% confidence interval (CI) of the observed data for survival, FEV1% predicted, and annual exacerbations (per patient per year [PPPY]. The model predicted 9.0 metres annual decline in Six Minute Walk Distance compared to ECLIPSE observed data of 5.7 metres decline. The model accurately predicted the TORCH placebo outcomes in at least two of the three annual time points within the 95%CI of the observed data for FEV decline and annual exacerbations PPPY.  The model over predicted survival by 8% (absolute) compared to TORCH observed data at year 3.   CONCLUSIONS: As expected, the model more accurately predicted the ECLIPSE observed outcomes in the internal validation exercise, than TORCH outcomes in the external validation.