A demodulation system based on linear array photo electric detector is researched and realized to acquire reflective spectrum data of fiber Bragg grating（ FBG） sensor. Due to limited number of sampling point of linear photo electric detector

to study impact of different number of datas on peak-searching precision of FBG reflective spectrum

based on the platform of virtual instrument Lab VIEW

three commonly used peak-searching algorithms are analyzed and compared in FBG sensor demodulation system

such as centroid method

Gaussian fitting algorithm and polynomial fitting algorithm; advantages and disadvantages of different number of datas are compared and analyzed. Experimental results show that using seven sampling points data on reflection wave produces the lowest error

and precision and stability of Gaussian fitting algorithm is proved to be better.