— This paper deals with the problem of blind carrier-frequency offset (CFO) estimation in OFDM systems based on offset quadrature amplitude modulation (OFDM/OQAM) with null subcarriers. Specifically, by assuming that the number of subcarriers is sufficiently large, the received signal is modeled as a complex Gaussian random vector (CGRV). Since the OFDM/OQAM signal results to be a noncircular (NC) process, by exploiting the generalized probability density function for NC-CGRVs, the unconditional maximum likelihood (ML) algorithm for CFO estimation in non dispersive channel is proposed. Moreover, a modified version of the unconditional ML CFO estimator is considered. The performance of the derived algorithms, assessed via computer simulation, is compared with that of recently proposed estimators exploiting the second-order cyclostationarity and with the Gaussian Cram´er-Rao bound.