Developing countries continue to grapple with the burden of infectious diseases in managing their public health. The growing use of technology and machine learning in healthcare has seen significant adoption of diagnostic and predictive ML models mostly in developed countries with enough supportive resources. For resource-constrained countries with substantial populations living in informal settlements, there has been a growing need for the integration of ML in their healthcare management. In response, we design a sustainable machine learning model for improving the prediction of infectious diseases in informal settlements. We strive to address limitations of existing predictive models, particularly their inability to effectively handle noisy, incomplete, and resource-constrained datasets common in slum environments. The research is handled in phases with this paper covering the first objective which aims at understanding the prevalence of infectious diseases and existing management strategies in informal settlements of Kibera, Mathare, and Mukuru kwa Njenga in Nairobi, Kenya. To achieve this, we adopt a mixed-methods research design to investigate the distribution and management of infectious diseases in selected areas. The subsequent phases of the research will evaluate existing machine learning datasets, and finally develop and evaluate a setting-specific ANN-based predictive model. The model will be trained using the backpropagation algorithm with stochastic gradient descent optimization and evaluated using standard machine learning performance metrics, including accuracy, precision, recall, F1-score, and confusion matrix analysis. In the first phase, the paper confirms high prevalence of Tuberculosis, Malaria, Cholera and HIV/AIDs, as well as existing management strategies adopted and the challenges in informal settlement. The proposed framework is expected to improve predictive accuracy by learning complex nonlinear relationships between clinical symptoms, environmental conditions, and socio-demographic factors. While we acknowledge that infectious diseases remain a serious health challenge in informal settlements, we believe that the proposed model provides a promising approach for developing sustainable, scalable, and cost-effective infectious disease prediction systems tailored to informal settlements. Future implementation should further validate the model using real-world clinical and community datasets, integrate it with existing health information systems, and continuously retrain it using updated epidemiological data. Policymakers should also support digital health infrastructure and standardized data collection to enhance predictive healthcare in informal settlements. This research contributes to the growing application of artificial intelligence in public health by demonstrating how optimized backpropagation-based neural networks can support sustainable infectious disease prediction in resource-constrained environments.