Frequency estimation algorithm for Minimum Shift Keying-Type(MSK-Type) signals is usually based on maximum likelihood theory. However

the algorithm is complex

and it is not practically realized in engineering. So the frequency synchronization for MSK-Type signal mostly adopts nonlinear methods such as delay-multiplication and 2P-Power algorithms. But these feedforward structure of non-data aided algorithm applies only to full response MSK-Type signal

and due to the nonlinear process of multiplying

these algorithms introduce large amounts of self-noise

so that it leads to poor synchronization performance. In order to solve the problem of lower accuracy and narrower application range and unpractical realization in engineering for the existing nonlinear algorithms

an improved algorithm based on delay-multiplication frequency synchronization algorithm is proposed in this paper. After different power nonlinear processing on the MSK-Type signal with partial response and full response

the improved algorithm eliminates the effects of self-noise in the delay-multiplication algorithm. Compared with the original algorithm

the estimated accuracy for the frequency offset is improved about 5dB

and at the same time

the application range is extended to the partial response MSK-Type signal. The improved algorithm enlarges the application range and it is easy to be implemented in engineering.