The invention relates to a fractional order ambiguity function algorithm based on fractional lower order statistics, and belongs to the technical field of computer applications. The algorithm comprises the following steps of step (1), proposing an instantaneous correlation function based on the fractional lower order statistics, and performing fractional Fourier transform on the instantaneous correlation function; and step (2), within a fractional order transform domain, realizing the joint estimation of cubic phase signal parameters by searching a peak point of an FLOS-FAF (the fractional ambiguity function based on the fractional lower order statistics) and a de-chirping method. Based on a relationship between the algorithm of the invention and a generalized signal to noise ratio, and a noise characteristic index, and compared with a fractional order ambiguity function algorithm, the algorithm of the invention can not only effectively suppress the interference of impulse noises, but also has a better estimation performance.