Accurate position determination of a wireless device by time-of-arrival (TOA) and angle-of-arrival (AOA) estimation of the line-of-sight (LOS) path with an array of collocated antennas, is important in various applications. In an indoor environment, there can be many multipath reflections, from different angles of arrivals; and, when the signal bandwidth is relatively narrow, the received signal from the LOS path is often obscured by the overlapping received multipath signals. In this case, accurate estimation of the LOS TOA and AOA is challenging. In this paper, we present a novel estimator for the LOS TOA and AOA in multipath conditions, by obtaining an approximate distribution for the received signals of all the antennas. We approximate as Gaussian the distribution of the received signals given the LOS TOA and AOA, as well as additional clusters’ mean AOAs; and subsequently derive a maximum likelihood estimator for the LOS TOA and AOA. We also further derive an estimation algorithm with reduced complexity. The performance of the LOS TOA and AOA estimation algorithm in challenging multipath conditions is evaluated. The results show, that the proposed estimator achieves a performance advantage compared to reference methods for sufficiently high SNR.