The increasing demand for electricity encourages the utilization of renewable energy such as solar panels. However, solar electricity production faces efficiency challenges as the panel output is highly dependent on light intensity and ambient temperature. This test aims to optimize the output power of solar panels using Maximum Power Point Tracking (MPPT) with Perturb and Observe (PO) method and light sensor-based prediction (LDR). The test was conducted using 100 Wp monocrystal solar panel with MPPT algorithm implemented through buck-boost converter and controlled by Arduino Mega 2560. The results showed that the MPPT system increased the output power efficiency by 16.13% compared to the non-MPPT system. Variation of light intensity from 0 to 10,400 lux resulted in an increase in voltage from 1.15V to 25V, with maximum power increasing from 0.023W to 13W, reaching an average of 800 LUX/W. Characterization of LDR resulted in a conversion factor of 7,761.194 LUX/LDR, enabling accurate prediction of MPPT values based on light intensity. Comparative analysis between the LDR and PO methods showed the LDR method reached a maximum power of 11.62W at 9,293.71 lux, while the PO method reached 12.51W at 8,500 lux, indicating comparable performance in optimizing solar panel output power.