This study aims to investigate the efficiency of employing IoT convolutional neural networks (CNNs) algorithms in smart chairs for sitting posture correction, comparing them with sensor fusion algorithm, in order to improve accuracy. In the context of smart chairs for sitting posture correction, the study used a comparison analysis between IoT CNN algorithms and sensor fusion algorithms. The execution of the IoT CNN method made use of deep learning frameworks, while the sensor fusion method combined data from several sensors. Using accuracy metrics, the two algorithms’ efficacy was evaluated. The results revealed that the IoT CNN algorithm illustrated higher accuracy in detecting and correcting sitting postures in comparison with sensor fusion algorithm. The CNN algorithm exhibited robustness in recognizing complex features of different postures, leading to more accurate corrections. In conclusion, the study emphasizes the effectiveness of employing IoT CNN algorithms in smart chairs for sitting posture correction.