Understanding the expression of neuropsychiatric risk-genes during early human brain development is crucial for unravelling the pathophysiology of mental health disorders (MHDs) and for developing potential therapeutic interventions. However, research on the prenatal brain transcriptome faces limitations, including limited access to fetal samples and the inability of short-read RNA sequencing technology to characterise full-length transcript isoforms. Here, we employed cutting-edge models, including human pluripotent stem cells (hPSCs)-derived cortical neurons and cerebral organoids, to recapitulate the early stages of human brain development. Leveraging the power of nanopore long-read RNA sequencing, we aimed to construct comprehensive and dynamic risk gene isoform-profiles that spanned up to six months of brain development. By integrating the obtained data, we revealed the isoform characteristics of seven well-established MHD risk genes at multiple timepoints throughout the developmental process. Among the 387 isoforms identified, 341 were novel, collectively constituting 55% of total gene expression, indicating the high diversity and prevalence of novel isoforms in developing brains. Additionally, the isoform expression trajectories of genes including CLCN3 and MAPT exhibited temporal shifts in isoform usage proportions, suggesting active regulation of corresponding molecular and cellular pathways at specific stages during corticogenesis. Furthermore, by comparing with the isoform profiles in post-mortem adult brains, we identified genes with large differences in isoform composition between developmental and adult samples, particularly MAPT. The isoform composition of MAPT in brain models is dominated by novel isoforms unique to development, which may hold pathophysiological significance in a neurodevelopmental context. This study underscores the unique nature of risk gene isoforms that participate in brain development and the possibility many developmental and pathological process are controlled by previously unidentified gene products. These findings will enable future research into their associations with MHD pathogenesis.