Lower Urinary Tract (LUT) dysfunctions, such as urinary incontinence or urinary retention, represent a clinically relevant group of conditions that significantly impact patients’ quality of life. Neuroprosthetic devices used for the treatment of the LUT dysfunctions are typically based on continuous open-loop stimulation paradigms, which may lead to voiding dysfunctions due to neural adaptation. In contrast, closed-loop stimulation strategies have demonstrated improved bladder capacity and voiding efficiency compared to continuous stimulation. This work has been developed within the framework of the Bladder Control project, which aims to overcome the limitations of open-loop approaches by introducing closed-loop neuroprosthetic systems. Such systems rely on the acquisition of electroneurographic (ENG) signals and on decoding strategies based on machine learning (ML) techniques to estimate bladder-filling states and apply nerve stimulation accordingly.