Wind energy is a potent yet freely available renewable energy. It is essential to estimate the wind speed (WS)precisely to makeaprecise estimation of wind power at wind power generating stations.Generally, the WS data is non-stationary. Wavelets have the potential to deal with the non-stationarilyindatasets. On the other hand, the prediction ability of primal least square support vector regression (PLSTSVR) has never been tested to best of our knowledge for WS prediction. Hence, in this work, wavelet kernel-based LSTSVR models are proposed for WS prediction. They are Morlet wavelet kernel LSTSVR and Mexican Hat wavelet kernel LSTSVR.HourlyWS data are collected from four different stations namely Chennai, Madurai, Salem and Tirunelveli in Tamil Nadu, India. The performance of the proposed models isevaluated using root mean square, mean absolute, symmetric mean absolute percentage, mean absolute scaled error and R2. The results of the proposed models are compared with twin support vector regression (TSVR), PLSTSVR and large-margin distribution machine-based regression (LDMR). Based on the results of the performance indicators, the performance of the proposed models is better when compared to other models.