With the intensiffcation of population aging and increasing burden of chronic diseases, the demand for vital signs monitoring is becoming increasingly urgent. A key challenge facing current non-contact detection technologies using millimeter wave (mmWave) radar is the low efffciency of multi-channel signal fusion in array radar systems based on equal weighting. To address this challenge, this paper proposes a vital sign enhancement detection method for multiple input and multiple output (MIMO) bio-radar, driven by multidimensional physiological characteristics, which overcomes traditional limitations through a two-stage fusion strategy. Stage 1: Enhanced Vital Sign Detection Using Single-Channel Signals Based on Physiological Characteristics. First, a chest wall multi-scattering point model is constructed. For single channel time-distance two-dimensional echo signals, effective range bins are selected based on the respiratory/cardiac physiological frequency band energy ratio, and the signal-to-noise ratio (SNR) of respiration/heart signals is enhanced using phase-aligned maximal ratio combining (MRC). Stage 2: Multi-Channel Fusion Based on Organ Radiation Spatial Distribution Characteristics. The spatial radiation characteristics of cardiopulmonary organs are introduced for the ffrst time as the theoretical foundation for SNR-based channel screening, channel attribute identiffcation, and multi-channel weighted fusion. Then, we propose a template matching method to extract respiratory rate (RR) and heart rate (HR) by adopting physical models of respiration and cardiac activities. The experimental results demonstrate the existence of the spatial distribution characteristics of organ radiation. In addition, we analyzed the impact of distance and state on the algorithm from these two aspects.