The investigation of biomolecular dynamics can reveal the most fundamental molecular mechanisms of cell function and disease formation. In recent years, single molecule tracking has shown the potential to revolutionize the study of biological systems. Recently we developed a real-time 3D single molecule tracking method known as 3D single-molecule active real-time tracking (3D-SMART), which uses active feedback to "lock-on" to single molecules in solution. 3D-SMART is capable of tracking single fluorophores in solution for several minutes at a time, two orders of magnitudes longer than previously reported methods, with photon limited temporal resolution. We demonstrate the application of 3D-SMART to track the free, 3D diffusion of DNA, RNA and proteins in the solution. Using two-color observation channels, 3D-SMART was also applied to observe the transcription of single DNA molecules in the solution by monitoring the real-time production of mRNA. Additionally, we also developed a deep learning-enhanced single-molecule spectrum imaging method, termed SpecGAN. In SpecGAN, the photon flux required to extract a single-molecule fluorescence spectrum can be reduced by 100 times, which enables two orders of magnitude higher temporal resolution compared to the conventional single-molecule spectrometer. With SpecGAN, the super-resolution spectrum image of the COS-7 membrane can be reconstructed with merely 12 000 frames of single-molecule localization images, which is almost half of the previously reported frame count for spectrally resolved super-resolution imaging. The combination of 3D-SMART and SpecGAN allows us to perform multi-dimensional single molecule tracking, which can reveal the 3D position dynamics and spectrum dynamics simultaneously. This multi-dimensional single molecule tracking method promises to be a powerful tool for capturing the biophysics dynamics of single biomolecules at high speeds and over 3D distances.