A positron emission tomography (PET) radioligand for imaging phosphodiesterase 4D (PDE4D) would benefit drug discovery and the investigation of neuropsychiatric disorders. The most promising radioligand to date, namely, [11C]T1650, has shown unstable quantification in humans. Structural elaboration of [11C]T1650 was therefore deemed necessary. High target affinity in the low nM range is usually required for successful PET radioligands. In our PDE4D PET radioligand development, we formulated and optimized an empirical equation (log[IC50 (nM)] = P1 + P2 + P3 + P4) that well described the relationship between binding affinity and empirically derived values (P1-P4) for the individual fragments in four subregions commonly composing each inhibitor (R2 = 0.988, n = 62). This equation was used to predict compounds that would have high inhibitory potency. Fourteen new compounds were obtained with IC50 of 0.3-10 nM. Finally, eight compounds were judged to be worthy of future radiolabeling and evaluation as PDE4D PET radioligands.