Changes in platelet count are associated with a variety of diseases and treatments. Measuring it gives better insight into the expected outcome. Our aim was to evaluate the accuracy of three methods for platelet count in hyperlipidemia samples. Sixty non-lipid whole bloods from 60 individuals were included: 20 in low platelet count group, 20 in medium platelet count group, and 20 in high platelet count group. Then, 400 μL plasma was exchanged with 400 μL fat emulsion. Platelet count was measured after replacement by three methods, impedance method (Plt-I), optical method (Plt-O), and fluorescence method (Plt-F). In the low platelet count group with fat emulsion plasma exchange, except for Plt-O, other methods showed the predefined acceptance criterion (± 10%) covered the mean bias and 95% CI of proportional bias (slope) which were obtained from Bland-Altman plot and Passing-Bablok algorithm, respectively. In medium and high platelet count group with fat emulsion plasma exchange, the predefined acceptance mean bias and criterion of 95% CI of proportional bias (slope) and intercept were met only in Plt-F. In the medium platelet count group, the mean bias was -1.600% and the 95%CI of slope and intercept were 1.000 (0.815 to 1.071) and -0.500 (-12.831 to 23.907), respectively. In the high platelet count group, the mean bias was -2.250% and the 95%CI of slope and intercept were 1.071 (0.974 to 1.225) and -33.8142 (-113.703 to 8.339), respectively. Our results show that the Plt-F can more accurately reflect the true platelet count of lipemia specimens compared with Plt-I or Plt-O.