Cardiotocography is a monitoring method that provides crucial and essential information on the condition of the fetus throughout both the time before birth and during labor. Modern advances in obstetrics allow for reliable machine learning methods for fetal heart rate categorization. The significance and interconnection of machine learning methods in illness diagnosis are progressively growing. This study uses a range of machine learning algorithms to forecast the well-being of the fetus based on the cardiotocographic (CTG) data. The health condition is categorized into three labels: normal, requiring assurance, and pathological. This research examines CTG characteristics to predict fetal health. This prediction is made using algorithms based on the Kernel Extreme Learning Machine (Ker_ELM). To improve the performance of the classifier, the voting process is employed with a preprocessing step to remove unwanted attributes. By integrating kernel methods into the Extreme Learning Machine framework, our model effectively captures non-linear patterns in fetal heart rate signals, leading to improved classification accuracy. Based on comparison data, all machine learning algorithms performed well. However, Ker_ELM stood out with an accuracy of 99.57%, precision of 98.56%, recall of 96.45%, and F1-score of 98.89%.