Islanding detection is an essential requirement of the modified power system scenario with increased penetration of DGs. A new detection technique based on wavelet energy entropy identification and active frequency drift confirmation approach has been proposed. Wavelet energy entropy is proposed as a solid indicator for possible islanding operation. As an active approach, frequency drift anti-islanding algorithm is proposed robustly shift the voltage frequency into the tripping window. The proposed method reduces time of detection leading to reliability. The technique also reduces non detection zone of passive method approximately to zero. These benefits of proposed method not only improve dependability of the detection but also overcome the power quality degradation issue of active method, as perturbation is not applied continuously in the system. The time taken by this dual level approach is much shorter than other hybrid methods as response time of wavelet entropy indicator is in ms and speed of active frequency drift is improved by adjusting parameters without any restriction on power quality degradation.