To　overcome　the　disadvantages　of　the　existing　control　algorithm　of　two-wheeled　robot(TWR),　the　adaptive　controller　was　designed　by　using　the　self-learning　parameter　optimization　algorithm.　The　controller　has　a　simple　structure　and　does　not　need　to　rely　on　the　accurate　mathematical　model,　and　the　optimal　control　parameters　can　be　obtained　through　learning　and　training　many　times.　The　controller　was　applied　to　the　balance　control　of　the　two-wheeled　robot,　and　was　compared　with　the　linear　quadratic　regulator　(LQR)　optimal　controller.　The　simulation　results　verify　the　correctness　and　validity　of　the　algorithm,　highlights　the　strong　robustness　and　bionics　habits.　The　controller　was　applied　to　the　physical　system　of　two-wheeled　robot,　and　achieved　good　control　effect.　©　2017,　Editorial　Department　of　Journal　of　Beijing　University　of　Technology.　All　right　reserved.