A selection detector in adaptive algorithm with weight constrained method for active noise control (ANC) is presented in this paper. Imposing weight constraints to the adaptation of the adaptive filter will effectively avoid over-updated phenomenon, such as a large movement of microphone or the sudden variations of the surrounding environment. It can be easily added in the adaptive procedure and will not burden heavily with complex computation. The selection of the weight constraint factor is prominent because a large weight constraint factor gives less constraint to the update of adaptive filter, whereas a small weight constraint factor may cause constraint of the update strictly. The selection detector is proposed to reduce the seeking time of weight constrained factors by imposing weight constraints to allocate the adaptive filter to effectively avoid over-updates phenomenon when there is a sudden large change in the original signal. From the selection detector block, these three weight constraint factors can be chosen appropriately to the minimum values of mean square error. The analysis gives useful information about preventing the adaptation from updating by the wrong learning and the robustness of the entire ANC system is improved in this paper. The results are verified to show the effectiveness through the computer simulations in this paper.