Time-dependent diffusion MRI (TDDMRI) is a powerful tool for probing tissue microstructure, but it faces challenges such as low signal-to-noise ratio (SNR) and resolution, motion-induced artifacts, and cerebrospinal fluid (CSF) contamination. This study aims to address these critical issues by developing a suite of novel 3D whole-brain TDDMRI techniques. First, to overcome low SNR and resolution, a 3D oscillating gradient preparation (OGprep) GRASE sequence was developed, which significantly improved image quality and acquisition efficiency compared to conventional 2D EPI methods. Second, to enable high-fidelity, whole-brain multi-segment acquisition, dual-navigator and subsequently 3D-navigator modules were integrated to correct for inter-segment phase errors caused by motion, complemented by GRAPPA parallel imaging to reduce scan time. Finally, to eliminate CSF signal contamination, an inversion recovery (IR) module was incorporated (3D IR-OGprep-GRASE), enhancing the accuracy of diffusion measurements in adjacent tissues like the hippocampus and in tumors. Validated on a clinical 3 T MRI scanner, these combined techniques successfully achieved high-resolution (1.5 mm isotropic) whole-brain TDDMRI, effectively suppressed artifacts, and provided more accurate microstructural parameter measurements than standard methods. This work establishes a robust and clinically viable imaging protocol that overcomes major limitations of TDDMRI, paving the way for its broader application in neuroscience and clinical diagnostics.