Advanced battery management is now essential for lifetime and safety due to the quick spread of lithium-ion batteries in electric cars and renewable energy systems. Static hardware thresholds are usually used by traditional battery management systems (BMS) to guard against overvoltage, undervoltage, and overcurrent. Although these static methods work well for simple defects, they frequently fail to anticipate compounding stress elements that cause thermal runaway or cell deterioration. The design and implementation of a sophisticated, Internet of Things-enabled BMS for a 3S (11.1V) Lithium-ion battery pack is shown in this study. It integrates real-time telemetry with an edge-computed Machine Learning (ML) algorithm. The system accomplishes high-fidelity sensing by using a 16-bit ADS1115 Analogue-to-Digital Converter and an ESP8266 microprocessor dynamic "Risk Score" is continually computed by a Logistic Regression model using voltage, current, temperature, State of Charge (SoC), and State of Health (SoH). The system has a multi-tiered protection procedure that includes a physical relay cutoff during critical stages and a nonblocking audio warning during medium-risk conditions. Additionally, the BMS transmits telemetry to the ThingSpeak cloud platform in its capacity as a smart-grid edge device. The dependability and remote observability of energy storage systems are greatly improved by this predictive method, as demonstrated by experimental data.