Abstract: With a limited number of exit temperature measurements of a gas turbine combustion chamber,it is most unlikely to take a measurement just at the correct position of the real hot spot.On the other hand,numerical prediction of the hot spot temperature so far is impracticable.In this paper,on the basis of principles of statistical analysis,a very simple and effective two dimensional optimization method is employed to process the multi measured temperature data taken at an annular combustion chamber exit under a specified operating condition.As expected,an optimum fitting Weibull equation is obtained,which can meet the preset accuracy requirement of evaluation by applying least square method,and then a calculation formula is derived,which can be used to ascertain the most probable exit peak temperature value with confidence.A numerical example shows that the most probable exit peak temperature(namely the real hot spot temperature)is higher to a certain extent than the maximum value measured.Some other potential applications of this method are also metioned very briefly.