Previously we developed a promising synthetic approach to complex polycyclic natural products, including aromatic polyketides derived from the type-II polyketide biosynthesis. The approach consists of three key steps; (1) cycloaddition or cyclocondensation of nitrile oxide for assembling the carbon skeleton having an isoxazole as a 1,3-diketone equivalent, (2) benzoin cyclization for stereoselective ring fusion with an angular hydroxy group, and (3) oxidation of isoxazolium salt for installing an additional angular hydroxy group in stereo selective manner. As a synthetic exercise to highlight the utility of this approach, the first total syntheses of the proposed structure of pleospdione (1) and its C3-epimer (3-epi-1) have been achieved. The synthesis has proved the efficiency of the strategy as an entry into polyketide-derived complex architectures. At the same time, it was revealed that neither 1 nor 3-epi-1 represent the true structure of the natural product by incongruity of the NMR spectra.