This paper proposes an adaptive algorithm for the combined acoustic feedback and noise cancellation in the binaural hearing aids. The proposed algorithm is based on dual microphones for feedback cancellation and the beamforming method for noise cancellation. The coefficients of feedback canceller are updated after subtracting the speech signal from the input signal by dual microphones. And the noise canceller reduces the noise signal in the residual signal excluding the speech by the beamforming method. Firstly, the feedback canceller operates to cancel the feedback signal in the microphone signal, and then the noise canceller operates to reduce the noise in the residual signal. Also, to assure the stable convergence of binaural hearing aids in the training mode, the coefficients of the left hearing aid are firstly updated, then the coefficients of the right hearing aid are updated. In the normal mode, the feedback and the noise canceller are operated without updating coefficients except an unstable case. To verify performances of the proposed algorithm, we analyzed its convergence behavior and simulated for real speech. From the results of simulations, it was proved that we can advance 14.43dB SFR(speech-to-feedback ratio) on average in the feedback canceller, 10.19dB SNR(speech-to-noise ratio) improvement on average in the noise canceller, in the case of applying the proposed algorithm.