Child and maternal mortality remain pressing global health concerns, particularly in low-resource settings. The United Nations has set ambitious targets to stop preventable deaths in just born babies and children under five years old, emphasizing the critical need for effective healthcare interventions. Maternal mortality, predominantly occurring in such settings, underscores the urgency to implement accessible and reliable screening methods for fetal health assessment. This paper focuses on leveraging Cardiotocograms (CTGs), a cost-effective and widely available tool, to classify fetal health into three categories: Normal, Suspect, and Pathological. An initial exploration of feature importance and selection done with ANOVA F-ratio, employed state-of-the-art machine learning algorithms to construct accurate classification models. Support Vector Machines (SVM), Random Forest Classifier, and Multi-Layer Perceptron were trained and optimized using grid and randomized search techniques with hyperparameter tuning. The highest achieved accuracy of 94.1% underscores the effectiveness of the proposed approach (FHCA- Fetal Health Classifier Algorithm) in fetal health classification. By accurately distinguishing between different states of fetal health, healthcare professionals can promptly intervene to prevent adverse outcomes for both the child and the mother. This study contributes to advancing prenatal care practices by demonstrating the utility of CTGs in enhancing fetal health monitoring. The findings highlight the potential of machine learning techniques to augment clinical decision-making and mitigate child and maternal mortality risks. Incorporating these strategies into healthcare systems, especially those with limited resources, has the potential to help achieve the Sustainable Development Goals outlined by the UN, ultimately resulting in the preservation of countless lives and the promotion of healthy communities across the globe.