The problem of estimating the fundamental frequency of a harmonic series arises in many areas of science and technology. For example, in vibration diagnosis, it is required to estimate the wear of bearings by the shift of the base of a harmonic series. In audio signal processing, this problem is associated with automatic instrument tuning. In speech synthesis, the fundamental frequency determines the pitch. In speech recognition, the frequency of the fundamental tone is an important information feature. In radio engineering, this problem is solved for the purpose of signal restoration, filtering, and decoding. In biomedical engineering, when analyzing a patient’s ECG, EEG, voice, or breathing, pathologies such as arrhythmia are diagnosed by the fundamental frequency. In the detection and classification of sea vessels, the most significant information criterion is the base of a propeller shaft–blade harmonic series. This paper proposes new approaches to estimating the fundamental frequency in high noise conditions. To reduce errors, the idea is to use the method of periodograms, filtering, autocorrelation, and the Hilbert transform. Note that in high noise conditions, the estimate of the fundamental frequency of a harmonic series is significantly improved by selecting optimal parameters: the size of the time window, filtering parameters, the spectrum interval for autocorrelation, and the number of autocorrelations.