Understanding of many physical phenomena in heat transfer requires both mathematical modelling and experimentation. A bridge between these is the study of inverse problems which can be classified as parameter or function estimation. Parameter estimation with actual data is emphasized for the determination of some thermal properties in solids. For this typical data, the number of parameters need not be large and the Gauss-Newton minimization method has advantages because it utilizes the individual sensitivity coefficients. These coefficients are used in the sequential estimation of the parameters, determination of confidence regions and design of optimal experiments.