In this paper, combination of the cross-Wigner distribution (XWD) and the Viterbi algorithm (VA) for the instantaneous frequency (IF) estimation of frequency modulated (FM) signals in high noise environments is proposed. The favourable properties of the VA, the IF reconstruction based on minimization of the path penalty functions, and the XWD, iterative accuracy improvement of the IF estimation, give hybrid IF estimator with improved accuracy for high noise environments