This paper is concerned with the development of a methodology for quality control and evaluation in compressor assembly lines. The tools Artificial Neural Networks (ANNs), Failure Mode and Effects Analysis (FMEA) and Fault Tree Analysis (FTA) are used. Based on this approach, it is proposed the analysis of the main failure modes in compressor assembly lines and the automatic identification of these cases with neural networks. The objective is the reduction of the number of compressors not assembled within the company recommended technical standards. This proposal aims at the extraction of a feature of a primitive signal acquired by installed sensors in the measurement panel, and at the classification of the signs of perfect or defective compressors with a Neural Network. The obtained results are compared with the current system of measurement with the purpose of evaluating the proposal. The accuracy index of the proposed model is between 97% and 100% of correctly identified patterns.