Boron phosphide (BP) is a (super)hard semiconductor constituted of light elements, which is promising for high demand applications at extreme conditions. The behavior of BP at high temperatures and pressures is of special interest but is also poorly understood because both experimental and conventional ab initio methods are restricted to studying refractory covalent materials. The use of machine learning interatomic potentials is a revolutionary trend that gives a unique opportunity for high-temperature study of materials with ab initio accuracy. We develop a deep machine learning potential (DP) for accurate atomistic simulations of the solid and liquid phases of BP as well as their transformations near the melting line. Our DP provides quantitative agreement with experimental and ab initio molecular dynamics data for structural and dynamic properties. DP-based simulations reveal that at ambient pressure, a tetrahedrally bonded cubic BP crystal melts into an open structure consisting of two interpenetrating sub-networks of boron and phosphorous with different structures. Structure transformations of BP melt under compressing are reflected by the evolution of low-pressure tetrahedral coordination to high-pressure octahedral coordination. The main contributions to structural changes at low pressures are made by the evolution of medium-range order in the B-subnetwork and, at high pressures, by the change of short-range order in the P-subnetwork. Such transformations exhibit an anomalous behavior of structural characteristics in the range of 12–15 GPa. DP-based simulations reveal that the Tm(P) curve develops a maximum at P ≈ 13 GPa, whereas experimental studies provide two separate branches of the melting curve, which demonstrate the opposite behavior. Analysis of the results obtained raises open issues in developing machine learning potentials for covalent materials and stimulates further experimental and theoretical studies of melting behavior in BP.