This paper presents a wavelet denoising method utilizing an enhanced threshold function to enhance the real-time monitoring of total organic carbon (TOC) in seawater. The conventional threshold denoising method has limitations when handling complex signals, often resulting in errors and residual noise during data reconstruction. This paper introduces an enhanced threshold function that addresses the limitations of the conventional method and reduces noise in high-frequency signals. The simulation results show that the signal-to-noise ratio (SNR) and minimum mean square error (MSE) of the signal with noise measured by the analyzer are 23.0370 and 0.042075. The SNR and MSE of the signal after noise reduction by the enhanced wavelet transform are 34.7860 and 0.0028127. In contrast, the values of the SNR and MSE of the traditional threshold noise reduction are between the noisy signal and the enhanced wavelet transform. Experimental data show that the enhanced threshold method retains useful signals more effectively than the traditional threshold method. In the long-term sea test data, the truth values of TOC monitoring results before and after denoising show a high consistency, and the Pearson correlation coefficient reaches 0.954, which verifies the stability and reliability of the method.