The classical multi-sensor anti-bias association algorithms require complex calculation procedure, can not be used in real-time condition, for instance the algorithms based on image matching and reference topology feature. On the basis of tracks distribution feature on sea surface and Cell-Average Constant False Alarm Rate (CA-CFAR) detection theory, a real-time anti-bias association algorithm for Automatic Identification System (AIS) and the data track of radar is proposed, named confidential-association algorithm, to make real-time radar systematic error registration and multi-sensor information fusion come true. Monte-Carlo simulation results show that the accuracy of confidential-association algorithm maintains on a high level on the sea-surface environment, it has the superiority of simply calculation procedure and substantially reduces run-time in comparison to current anti-bias association algorithms. Radar automatically registration technique based on confidential-association algorithm reduces average error of the measured data by nearly 90%.