The paper analyzes the existing localization methods in the context of non-cooperative identification. There are many advantages of such systems like usability, capacity increase, possibility of being hidden and undetected for security reasons, etc. The main problem is the poor quality of biometric data and the need for their rapid processing that requires special algorithms for processing of biometric data. The iris of an eye (iris) is currently the most promising data source for the non-cooperative systems. The iris can be scanned over a distance, has high information capacity, and remains unchanged with time. The most timeconsuming step besides the iris recognition is a localization process. The existing techniques of iris localization are selected and analyzed. Advantages and disadvantages of each technique are considered. It is shown that promising localization technique for non-cooperative systems is based on the active circuit technique. Despite its high computational complexity, the method allows to take into account a large number of factors related to the non-cooperative recognition. It is found that the success of the technique is determined by preliminary stages, for example, the correct specification of initial parameters.