To improve frequency estimation accuracy, a frequency estimation algorithm based on cross information fusion was proposed. The algorithm was suitable for signals of short duration and low signal-to-noise ratio (SNR), which are common in engineering. Firstly, several different signal groups were obtained by grouping multisegment signals according to the guidelines of combination. Secondly, rotation factors were obtained according to complementary information of multisegment signals in each signal group. Thirdly, the average cross information fusion spectrum was achieved by the arithmetic mean value of all the cross information fusion spectra, which were calculated according to rotation factors. Finally, the average cross information fusion spectrum was peak-picked to get the frequency estimator. Algorithm analysis and experimental results showed that compared with the existing algorithms, the proposed algorithm was superior in precision, robustness, and noise immunity.