Data security remains a major concern in the Internet of Things (IoT) landscape due to the inherent limitations in computational power, memory capacity, and energy availability of IoT devices. To address these challenges, lightweight encryption algorithms have emerged as alternatives to conventional cryptographic methods, aiming to balance performance and security. This study evaluates the effectiveness of five encryption algorithms—SIMON64/128, SPECK64/128, XTEA64/128, PRESENT64/128, and AES128—on IoT devices through experimental analysis of their security strength, execution time, CPU utilization, memory usage, and power efficiency. The experiments were conducted on a Raspberry Pi 3B+ using C-based implementations to emulate realistic IoT scenarios. The findings reveal that AES128 offers the strongest security characteristics, including the highest Avalanche Effect (39.29%) and Differential Resistance Score (6.76/10), but at the expense of significant resource consumption. In contrast, SIMON64/128 and SPECK64/128 deliver superior performance in terms of speed and resource efficiency, making them ideal for low-power environments, albeit with concerns about potential cryptographic backdoors. XTEA64/128 emerges as a practical compromise, delivering moderate security and low power consumption without known vulnerabilities. Based on these results, AES128 is suitable for high-capacity IoT platforms prioritizing strong encryption, while SIMON and SPECK are preferable for resource-constrained devices, with XTEA serving as a balanced alternative. This research contributes a comparative framework to guide the selection of encryption algorithms for IoT systems, ensuring an optimal trade-off between security and operational efficiency.