A monitoring experiment system of submarine cable vibration signal was set up based on Brillouin optical time domain analysis (BOTDA). The vibration signals of submarine cable under three working conditions of anchor smashing, scouring and friction were obtained. The TSA-VMD-MPE method was proposed to reduce the noise of vibration signal. The tunicate swarm algorithm (TSA) was applied to optimize decomposition level and penalty factor of variational mode decomposition (VMD) to obtain intrinsic mode functions (IMF) components. The correlation between the IMF and original signal, and the IMF variance contribution rate were considered comprehensively to determine the multi-scale permutation entropy (MPE) threshold to reduce the noise of the vibration experimental signal. The experimental results show that the signal-to-noise ratio of the three kinds of submarine cable vibration signals is improved by 12.0296dB on average. The proposed MPE threshold determination method was applied to EEMD and CEEMD algorithms, and good noise reduction effect was obtained.