Excimer laser micromachining was a key process for producing features by removing parts of some photoresist layers. One hurdle was that the lower substrate layer was easily attacked or damaged in laser ablation of the upper photoresist layer. To overcome this, a real-time monitoring scheme of laser micromachining process was proposed on the basis of taking acoustic emissions arising from pulsed laser–material interaction as feedback signals and employing a diagnostic program. It was showed that extracted features (e.g. the RMS values) from sensor signals were indicative of the process. In particular, the feature values tended to have an abrupt change when laser beam approached the interface of consecutive two layers. Relying on this characteristic, we proposed a fuzzy expert system as the core part of the diagnostic program. The effectiveness of the program was verified with tests.