Providing the security and robustness for the model is the crucial as the attackers are more for the machine learning and deep learning model due to increase in the usage of the Artificial intelligence in many applications. Hence this vulnerability to the model has drawn an attention of the researchers. Inducing the changes in the model causes the changes in the weights which in-turn causes the misbehaviours in the predication. The commonly used steps for the attacking are changes in maps, surface decision changes. This paper proposes the methodology called as Efficient Gradient Integrated Attack based on FGSM (EGIA). It also describes about the fundamental concepts in adversarial attack and its applications. The model is implemented in the python scripting languages and achieved the considerable attack on the existence model.