Tetrahydroprotoberberine alkaloids have shown interesting polypharmacological actions at dopamine receptors and are a unique template from which to mine novel molecules with dual selective actions at D1 and D3 receptors. Such compounds will be valuable to evaluate as anti-cocaine therapeutics. Towards that eventual goal, we engaged an SAR study in which a series of C9 alkoxy analogues of the D1/D2/D3 ligand (-)-stepholidine that possessed or lacked a C12 bromo functionality, were synthesized and evaluated for affinity at dopamine D1, D2 and D3 receptors. We found that the analogues are generally selective for the D1 receptor. Small n-alkoxy substituents (up to 4 carbons in length) were generally well tolerated for high D1 affinity but such groups reduced D3 affinity. In the case of C12 brominated analogues, C9 alkoxylation also had little effect on D1 affinity for the smaller alkoxy groups, but reduced D2 and D3 affinities significantly. C12 bromination tends to increase D1 receptor selectivity. A number of compounds were identified that retain affinity for D1 and D3 receptors but lack D2 receptor affinity. Among them, compound 22a was found to be a selective D1/D3 dual antagonist (Ki = 5.3 and 106 nM at D1 and D3 receptors). Docking studies performed on the analogues at the D3 receptor revealed a number of interactions that are important for affinity including a critical N - Asp110 salt bridge motif, H-bonds to Ser192 and Cys181 and hydrophobic interactions between the aryl rings and Phe106 and Phe345. The analogues adopt an orientation in which ring A is located in the orthosteric binding site while the C9 alkoxy substituents attached to ring D project into the secondary binding pocket of the D3 receptor.