To improve the prediction speed and accuracy in indoor localization， a novel algorithm based on hybrid clustering and LASSO is proposed. Coarse localizer is taken by clustering matching and LASSO theory is used for fine localization. Besides the traditional received signal strength （RSS） based clustering， a coordinate-based clustering method is also used aiming at reducing the error caused by wrong cluster match. The similarity of the RSS coverage vectors is used as the criterion of clustering and cluster matching to reduce the computational complexity. The algorithm of LASSO is applied to recover RSS signal from noisy measurements with reduced demand for power and memory. Experimental results indicate that the proposed algorithm leads to an improvement on the fine positioning accuracy and online complexity. An average positioning error of 1.73 m is achieved with grid spacing of 1.8 m.