Nanosatellites and small satellites use multiple compact electronic systems to execute their essential functions of power control and communication and data processing. The entire mission can be affected if electronic components fail because satellites become unrepairable after reaching orbit. Current satellite health monitoring systems depend on fixed threshold values which use telemetry data to monitor temperature and voltage levels. The methods which exist at present can identify problems only after actual damage has begun to occur. The current situation requires development of predictive maintenance methods which can detect early signs of equipment failure before actual breakdowns happen. This research introduces a predictive maintenance system which utilizes telemetry time-series data to monitor satellite electronic equipment. The study analyzes major system components through temperature measurements, voltage readings, current flow data and reset events from essential onboard electronic systems. The process starts with telemetry data cleaning and normalization while the data extraction process begins with essential feature identification. The machine learning models are developed to identify normal and degraded operational patterns which enable them to predict potential system failures. The study demonstrates that the proposed method outperforms traditional threshold-based monitoring systems through its enhanced ability to predict faults. The developed system is well suited for nanosatellite platforms because it requires low computational resources and can be implemented either onboard or at the ground station. This method will enable future satellite operations to perform autonomous health monitoring and extend mission duration while enhancing the dependability of nanosatellite networks.