This work investigates the estimation of hop rate, epoch time and the power spectrum of frequency-hopping (FH) signals which are of major interest in military and commercial applications. Multiple-hop autocorrelation (MHAC) processor is used to estimate the hop rate of FH signals in the presence of A WGN. During the hop time the single-hop autocorrelation (SHAC) processor can be used to estimate the epoch time. Previous work on both estimators assumes the signal power is known a prior while in practical applications it usually is not .The modified MHAC and SHAG estimators proposed here do not require prior knowledge of signal power. Instead, the signal power is first estimated and then utilized in the hop rate or epoch time estimation. The performance of the proposed estimators has been tested using computational experiments where the results are compared with known values of hop rate, normalized epoch time and signal power in the range 50-100 hops/sec.. 0.1-0.5 and l-20mW respectively. It was found that the estimates of hop rate for known and unknown signal power approach the same value for SNR≥ 0dB and the error in the estimation is less than 10%. Similar results were obtained for epoch time, while the error in the estimation of signal power did not exceed 5% for SNR≥0dB. For spectral estimation, three methods have been used to estimate the spectrum of FH signals: the periodogram (FFT) method, contracting mapping method (CMM) and maximum entropy method (MEM). The performance of these methods in AWGN is examined from the aspects of accuracy in location the frequency, vulnerability to noise level (SNR), speed and dependence of the resolution on the data record length. In the simulation, the frequency spacing is taken to be 25 KHz and three tested frequencies (30.025MHz, 60.3MHz and 80.4MHz) were considered. The computed power spectra for the noisy FH signal are presented. The aspects mentioned above decide the choice of the method. The periodogram, though less accurate than MEM and CMM at higher SNR, but is faster than MEM and more accurate than CMM at low SNR.