Recently we reported the pharmacological characterization of the 9,10-dihydropyrrolo[1,3]benzothiazepine derivative (S)-(+)-8 as a novel atypical antipsychotic agent. This compound had an optimum pK(i) 5-HT(2A)/D(2) ratio of 1.21 (pK(i) 5-HT(2A) = 8.83; pK(i) D(2) = 7.79). The lower D(2) receptor affinity of (S)-(+)-8 compared to its enantiomer was explained by the difficulty in reaching the conformation required to optimally fulfill the D(2) pharmacophore. With the aim of finding novel atypical antipsychotics we further investigated the core structure of (S)-(+)-8, synthesizing analogues with specific substituents; the structure-activity relationship (SAR) study was also expanded with the design and synthesis of other analogues characterized by a pyrrolo[2,1-b][1,3]benzothiazepine skeleton, substituted on the benzo-fused ring or on the pyrrole system. On the 9,10-dihydro analogues the substituents introduced on the pyrrole ring were detrimental to affinity for dopamine and for 5-HT(2A) receptors, but the introduction of a double bond at C-9/10 on the structure of (S)-(+)-8 led to a potent D(2)/5-HT(2A) receptor ligand with a typical binding profile (9f, pK(i) 5-HT(2A)/D(2) ratio of 1.01, log Y = 8.43). Then, to reduce D(2) receptor affinity and restore atypicality on unsaturated analogues, we exploited the effect of specific substitutions on the tricyclic system of 9f. Through a molecular modeling approach we generated a novel series of potential atypical antipsychotic agents, with optimized 5HT(2A)/D(2) receptor affinity ratios and that were easier to synthesize and purify than the reference compound (S)-(+)-8. A number of SAR trends were identified, and among the analogues synthesized and tested in binding assays, 9d and 9m were identified as the most interesting, giving atypical log Y scores respectively 4.98 and 3.18 (pK(i) 5-HT(2A)/D(2) ratios of 1.20 and 1.30, respectively). They had a multireceptor affinity profile and could be promising atypical agents. Compound 9d, whose synthesis is easier and whose binding profile is atypical (log Y score similar to that of olanzapine, 3.89), was selected for further biological investigation. Pharmacological and biochemical studies confirmed an atypical antipsychotic profile in vivo. The compound was active on conditioned avoidance response at 1.1 mg/kg, a dose 100-times lower than that required to cause catalepsy (ED(50) >90 mg/kg), it induced a negligible increase of prolactin serum levels after single and multiple doses, and antagonized the cognitive impairment induced by phencyclidine. In conclusion, the pharmacological profile of 9d proved better than clozapine and olanzapine, making this compound a potential clinical candidate.