Productivity of professional service is difficult to measure, due to the high degree of customization, variable throughput time, and high degree of labor intensity. Using the complex design engineering for large construction projects as an example, this study reviews the common professional service productivity measurement problem of determining surrogate measures of inputs and outputs. This research was sponsored by the construction industry, and required a team of academic and construction executives to work closely to develop a procedure and a comprehensive empirical model for measuring engineering design productivity. The model addresses the complexity of productivity estimation arising from the interactions among the multiple outputs and variability of labor hours, both requiring surrogate measures. The mathematical model was statistically estimated using data from a large number of design engineering organizations. The statistical results and the model were externally validated in several organizations for their pragmatic usefulness. The proposed methodology is applicable to those service industries where there are high degrees of output variability and highly skilled professionals working to achieve complex objectives.