A technique adapted from the guarded-comparative-longitudinal heat flow method was selected for the measurement of the thermal conductivity of a nuclear fuel compact over a temperature range characteristic of its usage. This technique fulfills the requirement for non-destructive measurement of the composite compact. Although numerous measurement systems have been created based on the guarded-comparative method, comprehensive systematic (bias) and measurement (precision) uncertainty associated with this technique have not been fully analyzed. In addition to the geometric effect in the bias error, which has been analyzed previously, this paper studies the working condition which is another potential error source. Using finite element analysis, this study showed the effect of these two types of error sources in the thermal conductivity measurement process and the limitations in the design selection of various parameters by considering their effect on the precision error. The results and conclusions provide valuable reference for designing and operating an experimental measurement system using this technique.