Fiber Bragg Grating (FBG) sensors are widely applied in aerospace engineering, structural health monitoring (SHM), biomedical instrumentation, and industrial process control due to their compact form, immunity to electromagnetic interference, and ability to function in extreme conditions. However, their precision is often compromised by temperature-induced Bragg wavelength shifts, resulting in cross-sensitivity between strain and temperature. This paper introduces an adaptive temperature compensation framework that integrates linear regression and polynomial modeling to address this challenge. Linear regression gives a simple solution to situations of near-linear temperature dependency whereas polynomial modeling includes the more complex nonlinear variation for a higher degree of precision. Through statistical verification with controlled laboratory readings, it is understood that the overall mean absolute error can be reduced by as high as 32% using polynomial fit to that of a linear regression. The possibility of implementing such models in the field through dynamic conditions is further indicated by implementing a simulation that simulates real sensing situations. The proposed approach is cost effective, little calibration is involved, and it is not complex in terms of hardware as traditional dual - FBG or reference sensor techniques. These results point to its promise to be highly accurate and stable in sensing over applications that vary in temperature.