Multibeam echosounders and sonars are increasingly used in fisheries acoustics for abundance estimation. Because of reduced side-lobe levels in the beam-array pattern, the new Simrad ME70 multibeam echosounder installed on board Ifremer’s RV "Thalassa" has been designed to allow improved detection of fish close to the seabed. To achieve this objective, precise and unambiguous detection of the water-bottom interface is required, which raises the issue of bottom detection, especially in the outer beams. The bottom-detection method implemented in the ME70 is based on the amplitude of the reverberated echo. Such an approach is efficient for vertical beams, but less accurate for beams with higher incidence angles, typically 30°–40° for the beam configurations used on RV "Thalassa", where the incidence angle, the beam opening, and the nature of the seabed contribute to weakening the backscattered signal. Therefore, the aim of this study was twofold. First, we proposed to improve the current bottom-detection method based on the amplitude of the echo. Thanks to the split-beam configuration being available for all beams of the ME70, we also proposed to use the phase information in the backscattered signals of the outer beams, as is more commonly done with multibeam systems dedicated to seabed mapping. Then, we set a Bayesian estimation framework that takes into account the spatial continuity between adjacent echoes, giving more robustness to the bottom estimation itself. Results using data collected at sea for various bottom types are presented here.