In order to suppress the influence from the outliers in the peak envelop extracted by peak detector, an azimuth estimation algorithm is proposed. The Huber function based on maximum likelihood criterion is utilized to suppress the error exceeding threshold and decrease the outliers' weight in the algorithm. Furthermore, double thresholds are adopted to improve Huber function, eliminate the bigger outliers' weight, compress the smaller weight and provide a basis for selecting the threshold parameters. This method does not need the statistics of noise, and can effectively suppress the influence of the outliers. Simulation results show that the accuracy of bearing parameter adopted by this algorithm is better than that of the ordinary Huber estimator and the least square algorithm. Under condition of the power SNR of 3 dB, the estimation error is less than 0.5° which meets the requirements of TACAN system .