The performance of solar panels is highly influenced by environmental conditions, where anomalies such as dirt accumulation or shading can significantly reduce efficiency. This research aims to design and build an Internet
of Things (IoT)-based anomaly detection system for solar panels using a polynomial regression method to predict normal performance. The system utilizes an ESP32 microcontroller conner light intensity sensor to collect real-time data. The collected data was processed using MATLAB to build a 2nd-order (quadratic) polynomial regression model that maps the relationship between light intensity (Lux) and current output (mA). This model was then embedded into the ESP32 to compare the predicted current with the actual measured current, flagging an anomaly if a negative deviation exceeds a 15% threshold. The results demonstrated a highly accurate regression model (R² = 0.9956). The system effectively detected anomalies in the 50% and 75% shading test scenarios and successfully sent notifications to the Blynk platform. This system proves to be an effective and low-cost solution for the proactive monitoring of solar panel conditions.