Finger Knuckle Print is a finger based biometric system. It is a model which is the most widespread biometric authentication for an individual. The performance of Finger Knuckle Print recognition algorithms mostly depends on different pose variations. Due to deformations on the finger-knuckle samples, the result of the pose variations leads to false accept rate and it reduces the robustness level. In this paper, an efficient method is introduce to solve the intra-class finger-knuckle pose variations, called the Kernel Intra-Class Finger-Knuckle Pose Density Assessment (‘K’ Intra-Class FKPDA) to improve the robustness of the biometric system. ‘K’ Intra-Class FKPDA method map the user’s finger-knuckle prints samples in high dimensional feature space using the kernel process. Markov Graphic Model Mapped finger-knuckle print samples on high dimensional feature space to calculate the density level and make out the intra-class pose variations. Bessel Thomson Finger-Knuckle Pose Variation filtering process identified intra-class pose variation and it is removed in ‘K’ Intra-Class FKPDA method. To reduce the delay time ‘K’ Intra-Class FKPDA method applies a linear filtering model. Finally, using ‘K’ Intra-Class FKPDA removed intra-class pose variations and finger-knuckle print is documented with increased robustness level. PolyU FKP Database evaluated Experimental results on 165 images show that the new method can improve the robustness factor appreciably compared with typical Finger Knuckle Print based recognition approaches reducing the filtering delay factor. Experiment is conducted on false acceptance rate, recognition rate, and filtering delay time.