Abstract: For multi-lane objects identification of vehicle adaptive cruise control, the coordinates of on car radar is transformed and invalid objects are discarded based on lane information. Radar objects were tracked by comparing the radar information of current sample period with the predicted value of previous period. Object identification process was accomplished by applying threshold, sliding sample window and nearest object selection in different phases of calculation. In the curve road, norm central angle method was adopted for position compensation. Experiments indicate that the identification of multi-lane objects and curve road compensation could be realized in a satisfied way and the requirements of ACC system could be fulfilled.