The sensorimotor recovery of the upper extremity is an underexplored field, making it complex to develop effective rehabilitation goals and personalized therapy. Predictive models can be crucial in planning rehabilitation strategies, enabling the rehabilitation team to create suitable treatment plans based on these predictions. This prospective cohort study aims to develop and validate data-driven models for predicting upper limb sensorimotor recovery using conventional statistical and machine learning techniques. A physiotherapist will measure the outcome at four different time points, with these outcomes being divided into independent and dependent variables. Newly diagnosed stroke patients admitted to the hospital will be screened for selection criteria, and potential participants will be invited to take part in the study. Those who agree to participate will be informed and asked to provide written consent. Baseline demographic characteristics such as age, gender, dominance, ethnicity, level of education, premorbid status, comorbidities, site and subtype of stroke, CT and MRI findings will be recorded. Assessments, including NIHSS, motricity index, grip, and pinch strength, will be conducted within 7 days of the stroke, along with kinematic analysis of the upper limb functional task. Follow-up sessions will assess dependent variables including motor functions, sensory perception, and quality of life. After data collection, significant factors contributing to predicting sensorimotor recovery will be identified, and a prediction model will be developed and validated.