Filterbank Multicarrier (FBMC) modulation is a potential candidate for cognitive radio (CR) wireless networks due to its high spectrum efficiency and low interference among neighboring bands. Within the CR network it is important that the base station (BS) can localize the mobile station (MS) to better distribute radio resources. While most research has focused on interference within the FBMC network, We present a novel technique based on direct maximum likelihood (ML) position estimation which uses previously measured channel statistics to estimate MS locations by combining received signals from several BS. Our previous work has shown that this kind of direct ML estimation method outperforms conventional methods based on TOA estimation in Orthogonal Frequency Division Multiplexing (OFDM) systems. Here we demonstrate how this technique can be adapted to the FBMC system.