Aiming at the spectrum leakage that exists in fast Fourier transform (FFT),several ratio frequency estimation algorithms that widely used by engineering signal processing are deeply studied.Firstly,the principles and features of these algorithms are introduced,and the influences of three parameters on frequency estimation accuracy are concretely analyzed.These parameters are signal to noise (SNR),cycles covered in record and the distance from the DFT output peak to the true frequency in bin widths.Then,related simulations are carried out to analyze and compare the performance of these frequency correction methods.The simulation results show that the frequency estimation accuracy has lower limit.When the number of the waves in the tone is small,the accuracy of these methods is poor,but can be improved observably by increasing the waves for the most methods.When the number of the waves in the tone is large,most methods have high and stable accuracy.The performance of the Rife-Jane’s method and Vetterling’s method can be improved by using Hanning-tapered data.Finally,considering specific engineering environment,some rules and principles about how to choose appropriate algorithm are provided.