Barium promoted ruthenium catalysts deposited on the boron nitride supports were characterised (XRD, O 2 and CO chemisorption) and tested in NH 3 synthesis. Prior to use, the raw BN materials marked as BNS (Starck, 96 m 2 /g) and HCV (Advanced Ceramics Corporation Cleveland USA, 40 m 2 /g) were heated in an ammonia stream at 700-800°C for 120 h. As a result, the oxygen content was reduced from 7.0 at% (BNS) to 3.5 at% (BNS NH3 ) and from 3.8 to 2.7 at% (HCV NH3 ), as evidenced by XPS. The kinetic studies of NH 3 synthesis (63 or 90 bar; H 2 :N 2 = 3:1) revealed that the catalysts based on the modified supports were more active, respectively, than those derived from starting nitrides, the difference being especially pronounced in the case of BNS and BNS NH3 . Studies of the catalysts activation have shown, in turn, that the stabilisation in a H 2 :N 2 = 3:1 mixture at 1 bar is very slow, i.e. the reaction rate increases slowly versus time on stream even at a high temperature of 550 -600°C. Stabilisation is faster and the NH 3 synthesis rates are higher when the activation is performed with an ammonia rich mixture (10% NH 3 in H 2 :N 2 = 3:1) flowing under high pressure of 90 bar. It is suggested that boron oxide (an impurity) acts as a deactivating agent for Ba-Ru/BN and that the reaction between NH 3 and B 2 O 3 (B 2 O 3 + 2NH 3 = 2BN + 3H 2 O) is responsible for the activity increase. A poisoning mechanism of B 2 O 3 is discussed.