A system to support an incorrectly typed input keyword search in Thai language is proposed in this work. Segmentation and similarity measure algorithms are employed to enhance the traditional keyword search engine. The average of six similarity measure algorithms including Levenshtein, Overlap (bi-gram), Overlap (tri-gram), Jaccard, Dice (bi-gram), and Dice (tri-gram). The prototype is tested by 93 subjects including both native and non-native Phuket subjects. Top twenty-five Phuket attraction names are used as the data set. The experimental results show that the proposed system can improve the efficiency of the original search from 54.4% to 91.6% while the execution time of the extra steps can be negligible. Moreover, Bi-gram algorithms seem to outperform their Tri-gram counterpaths in this experiment and Jaccard seems to be outperformed by other similarity measure algorithms.