Artificial Intelligence (AI) has emerged as a promising solution in digital health, clinical decision support, and health informatics, enabling more informed and accurate patient outcomes. Numerous medical disciplines have successfully adopted Deep Neural Networks (DNNs) for the classification and diagnosis of various pathologies. Among these, gynecology and obstetrics represent critical fields where even minor physiological irregularities can lead to complications that may endanger both the fetus and the mother. Continuous monitoring systems are therefore essential to assess fetal development and detect potential risks. One such system is the cardiotocograph (CTG), a medical device that records the fetal heart rate (FHR) and uterine contractions (UC) of the mother. However, the signals obtained from CTG measurements are often complex and difficult to interpret, which can delay timely clinical interventions. Although several automated classification methods have been proposed in recent years, their predictive performance remains suboptimal. In this paper, we propose a hybrid classification model that integrates an AlexNet-based convolutional neural network for feature extraction with an Error-Correcting Output Codes (ECOC) framework employing Support Vector Machines (SVMs) as base learners for multiclass classification. The model was trained and evaluated using the open-source UCI Cardiotocography dataset. Experimental results demonstrate that the proposed architecture can offer a model that achieved superior performance compared to existing methods, attaining an accuracy of 95.18% and an F1-score of 95.01%, indicating its strong potential for clinical implementation in fetal health monitoring systems.