Diagnostic technology for photovoltaic (PV) systems was developed, using the learning method to take each site's conditions into account. This technology employs diagnostic criteria databases to analyze data acquired from the PV systems. These criteria are updated monthly for each site using analyzed data. To check the shadows on the PV modules and pyranometer, the sophisticated verification method was also applied to this technology. After the diagnosis, a basket method provides maintenance advice for the PV systems. Based on the results of precise diagnoses, this expert system offers quick and proper maintenance advice within a few minutes. This technology is highly useful, because it greatly simplifies the servicing and maintenance of PV systems. (C) 2002 Elsevier Science B.V. All rights reserved.