To improve the accuracy of the transient temperature detection system, Transient temperature inversion processing algorithms is proposed based on spectrum analysis of speckle pattern interferometry. The interference fringes are formed by speckle interferometry in the system. For transient temperature changes cause the material strain, the speckle interference pattern changes. The interference fringes on the measured surface are obtained by the area array CCD collection before and after deformation. The corresponding spectrum density function will changes by the changes of transient temperature, and the amplitude changes of center wavelength is inverted by the speckle pattern interferometry. Through detecting and calculating the ratio of the amplitude of the center wavelength, the transient temperature can be obtained by spectrum analysis. In the analysis and calculation for the function of transient temperature and material strain, material strain and interference fringes, derived the amplitude and phase function of the transient temperature change and interference fringes, so as to provide the necessary conditions for detecting the use of spectral density function temperature. 660 nm laser diode, and SI6600 type area CCD detector is used to obtain the speckle pattern interference fringes, system extracts the amplitude ratio of the central wavelength from spectral distribution function, and by calculation and calibration the detection accuracyoftemperature can be achieved of 2. Compared to traditional ±℃methods of direct detection forinterference fringes changes, the newmethodsimprove the accuracy by nearly an order of magnitude. It is more accuratedetection, betteruniformity and betterstability.