Abstract: In single-photon lidar for point target detection, the extremely low signal-to-noise ratio (SNR) of echo signals and their susceptibility to noise interference make target detection and ranging accuracy difficult to guarantee. To enhance the SNR and reduce the false alarm rate, this paper proposes a denoising algorithm combining particle swarm optimization (PSO) and variational mode decomposition (VMD). The algorithm introduces the PSO into the VMD process to optimize its key decomposition parameters, overcoming the limitations of traditional VMD, which relies on manual empirical parameter setting, resulting in insufficient universality and suboptimal parameters. Based on the spectral characteristics of single-photon lidar echo signals, the algorithm selects signal smoothness as the fitness function for PSO and sets the kurtosis coefficient as the mode selection criterion, thereby achieving effective separation of signal and noise. Experimental results demonstrate that the proposed denoising method achieves a root mean square error (RMSE) as low as 2.426 5. It not only improves the peak signal-to-noise ratio (PSNR) of the echo signal by 5.39 dB but also effectively retains the waveform characteristics of the original signal, significantly reducing the target false alarm rate by up to 87.1%. Compared to traditional denoising algorithms such as wavelet transform, empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), and VMD, the proposed algorithm achieves significant improvements in denoising performance.