This research is extension of previous research on method of Collaborative and Non-Collaborative Dynamic Path Prediction Algorithm for Mobile Agents Collision Detection withDynamic Obstacles in 3D Space. In this research the extension of the algorithm for dynamic collaborative path prediction for mobile agents is proposed and two important issues Coefficient of agility and Minimal sampling frequency are analyzed. Those two terms are proposed in previous research for which background and introduction is explained in this paper. Solving Coefficient of agility and minimal sampling frequency issues allows system designers and developers to implement the method. This method allows full decentralization of collision detection which allows many advantages from minimizing of network traffic to simplifying of inclusion of additional agents in relevant space. Implementation of the algorithm will be low resource consuming allowing mobile agents to free resources for additional tasks.