The internal standard methods are known to compensate for the errors from sample preparation and injection into an analytical instrument. However, recent HPLC apparatuses have injectors of excellent repeatability and it is dubious whether the cancellation of injection error can lead to substantial improvement in the precision of analysis. This paper answers the above question experimentally and theoretically. The HPLC analysis of butylscopolamine bromide is taken as an example. The relative standard deviations (RSD) of measurements in the internal standard method and absolute calibration curve method are compared and the advantages of these methods are discussed. The measurement RSD is shown to be well estimated by the (function of mutual information) (FUMI) theory without repeating measurements. This report also demonstrates simple equations for calculating the measurement RSD at an arbitrary concentration of analyte and for selecting the better method between the internal standard method and absolute calibration curve method under specific experimental conditions.