how to perform gmapping with two laser sensors attach on front and rear place of the robot ?

edit

You would have to merge the scans prior to input into most all SLAM libraries in ROS. Whether you use that package or another, somehow you need to preprocess it.

I am not sure about @stevemacenski's answer. Looking at the source code of gmapping it seems to be "smart" about it.

As long as your laser sources have a correct tf frame_id and your tf tree is properly set up, gmapping should be able to handle both lasers. You would just need to set up both laser nodes to output messages to gmapping's scan topic.

I have never tried it myself, so it might fail miserably. Let me know how it goes if you give it a try.