Adenosine deaminase (ADA) is a key enzyme of adenosine metabolism. There are currently various kits and systems available for ADA measurement, and all yield variable results. This study optimized a reference measurement procedure (RMP) for serum ADA for the standardization of routine methods. ADA coupled with purine-nucleoside-phosphorylase, xanthine-oxidase and peroxidase was selected as the basic method and was optimized using Response Surface Methodology. Then the performance was validated and the results were compared after replication by 3 other reference laboratories. A reference interval was also developed. In addition, this optimized method was applied to calibrate a routine system. The intra-assay precision was 0.44% at both concentrations of 29.8 and 100.4 U/L, and inter-assay precision was 1.01% and 0.95% at 30.1 and 100.3 U/L, respectively. The linearity was up to 351.9 U/L (R-2 = 0.9998), with no significant interference or carryover ( < 5%). A Comparison among 4 reference laboratories showed good reproducibility (R-2 >= 0.9975). The procedure proved valid for a reference interval of 11.7-38.5 U/L. The mean relative deviation for a routine system was - 55.9% and - 3.7% before and after calibration. This candidate RMP for serum ADA can potentially be used for standardization of clinical systems.