Effective biomedical waste (BMW) management is essential, especially in metropolitan areas. Many cities face difficulties with BMW Management leading to environmental contamination and health risks. This situation demands advanced strategies, such as predictive analysis and deep learning models, to improve waste handling and processing while addressing current challenges and the lack of awareness among healthcare practitioners. This paper analyzes biomedical and yellow-coded waste type generation in healthcare systems using deep learning modeling techniques, particularly artificial neural networks (ANN) and Gaussian Process Regression models (GPR). The validity of the inputs was confirmed using sensitivity analysis to develop several models. Based on the statistical analysis, it was revealed that yellow-coded waste which includes human and animal anatomical waste, soiled waste, medicine waste, contaminated linen and beddings, and laboratory waste had a very high positive frequency of occurrence (0.98) with the total BMW. There is very high prediction accuracy together with a low value of forecast errors observed in the developed models in terms of targeted outputs with R2 values of 0.99 in both the two models, whereas the GPR model outperforms ANN with minimum MSE values of 0.00003. Through applying these high-performing models, healthcare waste management can be significantly improved by ensuring proper planning, resource allocation, waste segregation, treatment, and disposal, as well as improved safety within a healthcare facility and the overall reduction of the negative effects on the environment. The study highlights the importance of advanced modeling techniques in improving BMW management and offers practical implications for policy development and environmental management strategies.