Recently, polymer electrolyte fuel cell (PEFC) begins to be used in a lot of fields such as automotive industry and power generation etc. Therefore our laboratory developed the simple maintenance equipment which can diagnose a performance degradation of PEFC instantaneously by analyzing transient response at a current interruption. However, because it was not able to diagnose some operating conditions, we determined that it was necessary to improve the diagnostic algorithm. Therefore we developed a new algorithm which sequentially approximates diagnostic parameters by the on-line algorithm of the method of least squares. The adequacy of a new algorithm was verified using a PEFC single cell of 25 cm 2 applied about two Membrane Electrode Assemblies (MEAs) ((a) 10 mg, (b) 7.5 mg) with different amount of platinum. As a result, we judged that this algorithm was able to apply to the new diagnostics because it distinguished the difference of two MEAs in each parameter.