Excimer laser ablation is used for microvia formation in the microelectronics packaging industry. With continuing advancement of laser systems, there is an increasing need to offset capital equipment investment and lower equipment downtime. This paper presents a neuro-fuzzy methodology for in-line failure detection and diagnosis of the excimer laser ablation process. Response data originating directly from laser tool sensors and the characterization of microvias were used as failure symptoms for potential deviations in four laser system parameters from their corresponding baseline values. The response characteristics consist of via diameter, via wall angle, and via resistance. Resistance measurements on copper deposited in the ablated vias were performed to characterize the degree to which debris remaining inside the vias affected quality. The laser system parameters include laser fluence, shot frequency, number of pulses, and helium pressure flow. The adaptive neuro-fuzzy inference system (ANFIS) was trained and subsequently validated for its capability in evidential reasoning using the data collected. Results indicated only a single false alarm occurred in 19 possible failure detection scenarios. In failure diagnosis, a single false alarm and a single missed alarm occurred.