This survey attempts to provide a detailed overview of the previous work on Uyghur on-line handwritten character recognition technology and mainly focuses on investigating the feature extraction algorithms that have been presented or applied to Uyghur characters. Statistical, structural and combined-hybrid approaches for feature extraction technology for Uyghur characters are going to be explored. Besides the popular feature extraction methods which perhaps are granted to be taken into consideration for all letter kinds, the study on Uyghur handwritten recognition contributed new features and extraction algorithms for pattern recognition. Summarization on observing methods and extracted features from the published work and suggestions for further attempts would be valuable reference for the up-coming research activities on finding more efficient features.