Successful disease modifying drug development for Alzheimer’s disease (AD) has hit a roadblock with the recent failures of amyloid-based therapies, highlighting the translational disconnect between preclinical animal models and clinical outcome. Although disease modifying therapies are the Holy Grail to pursue, symptomatic therapies addressing cognitive and neuropsychiatric aspects of the disease are also extremely important for the quality of life of patients and caregivers. Despite the fact that neuropsychiatric problems in Alzheimer patients are the major driver for costs associated with institutionalization, no good preclinical animal models with predictive validity have been documented. We propose a combination of quantitative systems pharmacology (QSP), phenotypic screening and preclinical animal models as a novel strategy for addressing the bottleneck in both cognitive and neuropsychiatric drug discovery and development for AD. Preclinical animal models such as transgene rats documenting changes in neurotransmitters with tau and amyloid pathology will provide key information that together with human imaging, pathology and clinical data will inform the virtual patient model. In this way QSP modeling can partially overcome the translational disconnect and reduce the attrition of drug programs in the clinical setting. This approach is different from target driven drug discovery as it aims to restore emergent properties of the networks and therefore likely will identify multitarget drugs. We review examples on how this hybrid humanized QSP approach has been helpful in predicting clinical outcomes in schizophrenia treatment and cognitive impairment in AD and expand on how this strategy could be applied to neuropsychiatric symptoms in dementia. We believe such an innovative approach when used carefully could change the Research and Development paradigm for symptomatic treatment in AD.