This paper presents an application of a Fuzzy Logic based Model for audiovisual speech recognition. The speech signal is analyzed with the Perceptual Linear Predictive Analysis and the speaker's face image is processed to extract some lip-opening parameters. Two sets of fuzzy filters are trained to extract some features from both acoustic and visual paths. These features are integrated by a second level of processing based on a feedforward neural network. Results obtained with a Signal to Noise Ratio between 0 and 18 dB are in the range of 63-95% and are compared to the Kohonen feature extractor.