In our previous publications, the response of perfluorinated (PF) graded index (GI) POFs (62.5/750, 62.5/490m) to bending, tensile loading, and cyclic loading was investigated. The results showed that Cytop-1 (62.5/750m) was more appropriate to be used as an optical fiber sensor for automotive seat occupancy sensing relative to Cytop-2 (62.5/490m). In this study, a textile-based optical fiber sensor was designed and the effect of automotive seat covering including face material and foam backing on a sensor's performance was analyzed. The pressure interval under which the proposed POF sensor design could perform well was found to be between 0.18 and 0.21N/cm(2), where PF GI POF (62.5/750m) was used as the POF material. The responses of the sensor in this interval were observed to be accurate and reproducible. The face fabric structure and the thickness of foam backing were not found to be significant factors to change the sensor response. Artificial neural network (ANN) was used for data analysis, and Qwiknet (version 2.23) software was used to develop ANNs. According to the results of Qwiknet, the prediction performances for training and testing data-sets were 75 and 83.33%, respectively.