Cerebral Palsy (CP) is one of the most common neurological disorders, primarily affecting children at an early age and leading to significant locomotor disabilities. Early diagnosis in premature infants and high-risk neonates is essential for timely intervention, enabling better rehabilitation outcomes. Recent research in areas of machine learning (ML) and deep learning (DL) has facilitated the development of automated diagnostic tools for CP using video and image datasets. In this work, we develop a novel system for CP detection using real-time live-streaming videos. The system integrates OpenCV for real-time image preprocessing, MediaPipe for extracting skeletal movement features, and Long Short-Term Memory (LSTM) networks for sequential motion analysis. By analyzing temporal movement patterns, the system effectively identifies abnormalities indicative of CP. Experimental evaluation demonstrates a high accuracy of 97%, underscoring the potential of this approach for reliable and efficient diagnosis. The proposed framework can be used as an assistive tool for healthcare professionals, offering an automated, non-invasive, and scalable solution for CP screening. By reducing reliance on manual observation and ensuring consistent evaluation, this system can enhance diagnostic precision and support early intervention strategies. Future work includes validation on larger and more diverse datasets and addressing real-world deployment challenges to optimize performance across various clinical settings. The results highlight the transformative potential of integrating real-time video analysis with advanced ML techniques in pediatric healthcare.