You can find the displacement from accelerometer output data (acceleration) by calculating the differentiation  to get the velocity then repeat the same with velocity to get the displacement using any software such as Origin pro.

A simple solution as suggested by Abdel-Rahman works in theory but is complicated by signal drift and noise. Errors will exponentially increase when double-integrating the acceleration signal. As a result, the calculated displacement will be highly unreliable. Luckily, more sophisticated methods exist. One approach that has shown to work particularly well involves periodically applying a zero-velocity update in the integration process. However, this requires knowledge about an instance during which the sensor will be stationary. If your question is related to walking or running, I suggest reading up on Pedestrian Dead Reckoning and ZUPT algorithms. Here are some helpful resources:

Angermann, Michael, et al. "A high precision reference data set for pedestrian navigation using foot-mounted inertial sensors." Indoor Positioning and Indoor Navigation (IPIN), 2010 International Conference on. IEEE, 2010.

Fischer, Carl; Talkad Sukumar, Poorna; Hazas, Mike "Tutorial: Implementing a Pedestrian Tracker Using Inertial Sensors", Pervasive Computing, IEEE, On page(s): 17 - 27 Volume: 12, Issue: 2, April-June 2013

@Abdel-Rahman Akl 

@Clemens Markus Brahms: Thanks both of you. I will try to implement the suggestions and get back to you.

Thanks Brahms, yes I mean calculate the integration as a simple way to get the displacement, but definitely I agree with your opinion. So, we can use any software for processing and filtering data to decrease the noise. Many software can do these processes such as Origin pro, SigView, or MatLab.

Regards, 

Integrate the kinematic and dynamic equations.