This paper proposes an optimal technique for total harmonics distortion detection and estimation for smart meter by applying Fast Fourier Transform (FFT) methodology to detect harmonic distortion. Design considers the precision of constant frequency and inconstant frequent including unbalance voltage in order to enhance efficiency and mitigate harmonic distortion value. These results show that implemented algorithm can detect total harmonic distortion precisely with maximum 0.058% error. However, usually, the accurate of Total Harmonic Distortion estimation will depend on the sampling rate of input signal. The most significant operating key of the Smart Meter is to accurately and rapidly analyses the signal. Therefore, this paper aims to find optimum sampling rate between 12.8 kHz and 25.6 kHz. From the result, the optimum sampling is 12.8 kHz. Because result of %THD value we're close in every cases but at sampling rate 12.8 kHz used in less computation time than to reduce memory use and processing time.