This study investigated the wear of the valve and seat insert seating faces. A tester, an exhaust valve and a seat insert were used. Test variables were cycle numbers ($2{ imes}10^6,;4{ imes}10^6,;6{ imes}10^6;and;8{ imes}10^6$) and Hz (10Hz and 25Hz). The other test conditions such as temperature ($350^{circ}C$), fuel (LPG) and load (1960N) were fixed. The 10Hz tests indicated that the average Rmax of the valve increased at the rate of $7.76{mu}m/10^6$ cycles starting from $29.42{mu}m$ at the $2{ imes}10^6$ cycles and that of the seat insert increased at the rate of $8.57{mu}m/10^6$ cycles starting from $34.19{mu}m$ at the $2{ imes}10^6$ cycles. The 25Hz tests indicated that the average Rmax of the valve increased at the rate of $1.58{mu}m/10^6$ cycles starting from $74.2{mu}m$ at the $2{ imes}10^6$ cycles and that of the seat insert increased at the rate of $1.25{mu}m/10^6$ cycles starting from $83.95{mu}m$ at the $2{ imes}10^6$ cycles. The tribochemical reaction product covered the two seating faces, preventing the wear of the seating faces. As cycle numbers became greater, the average Rmax of the seating faces became greater, but the increase rate varied significantly depending on the Hz. The wear mechanism of the two faces was investigated through the tribochemical reaction.