Men`s artistic gymnastics consist of six apparatus. The last apparatus in the Olympic order is a horizontal bar, which due to the complex, virtuos and demanding elements that gymnasts perform has the epitome of royal discipline. Dismounts are a group of elements which gymnasts use to finish their routine and which therefore affect the final score and the competition rank. In this paper, all horizontal bar double salto backward stretched dismounts with rotations along longitudinal axis that are currently in the FIG Code of Points were analysed. The aim of the paper was to determine the kinematic requirement of the horizontal bar dismounts performance double stretched somersault backward with different amount of rotations around longitudinal axis of the gymnast and whether there are and which are the variables that differ dismounts between different turns around the longitudinal axis of the gymnast. Data collection took place during the official 2016 FIG World Challenge Cup tournament in Osijek. A sample was a total of 12 elite gymnasts (18-29 y.o., 159-183 cm, 55-80 kg), representatives of the national team from 10 different countries. For the purpose of kinematic analysis, 15 different dismounts were selected, divided into three groups by five dismounts, according to the rotations around the longitudinal axis (360°, 720°, 1080°). Data processing was performed according to the APAS procedures (Ariel Performance Analysis System). Five positions were defined in which a total of 128 variables were analyzed by the SPSS (Statistical Package for the Social Sciences) program package. The results confirmed the established research hypothesis that different amount of rotation around the longitudinal axis are characterized by a specific relationship between biomechanical variables and parameters defining the trajectory of the flight and initiating and controlling the rotation around the transverse and longitudinal axis and the hypothesis that the group of dismounts classified according to the number of turns around the longitudinal axis differ according to the parameters within certain variables. Statistical data processing has shown that in certain positions there are differences in parameters within certain variables and the difference between dismounts group in a total of 74 variables. In five specific positions of the body, variables have differ dismounts with 1 and 2 turns around the longitudinal axis in total of 14 times, with 2 and 3 turns 60 times and with 1 and 3 turns 84 times. The results showed that for performance of double stretched somersault backward dismounts with 3 turns around the longitudinal axis it`s necessary to significantly increase the values of parameters of multiple biomechanical variables that condition a safe and successful execution with respect to dismounts with 1 or 2 turns that interact with each other and with respect to dismounts with 3 turns in a smaller number of variables and in a smaller ratio. The information from this research suggests the need to achieve certain biomechanical parameters while learning the dismount double stretched somersault backward in order to pre-prepare the conditions for a successful upgrading of a double stretched somersault backward dismount by adding 1, 2, and then 3 turns around the longitudinal axis. In addition to help in the training process, the results indicated the need to consider assigning a higher difficulty value to the dismount Fedorchenko in the FIG Code of Points.