This paper proposes a user-centric access point selection scheme to minimize the number of unnecessary reconfigurations of cognitive terminal. In this scheme, number of terminal reconfigurations is defined based on the dwell-time of terminal which is estimated from user's mobility and coverage area of the access point (AP). Each terminal selects one AP from a candidate list that is generated based on user's preferences. Information needed for the selection scheme is obtained from both network-side and terminal-side entities. Four different algorithms are investigated in order to find the best combination of dwell-time and other parameter such as load of APs, to minimize the number of reconfigurations and to achieve load balancing for Radio Access Networks simultaneously.