Current validation and verification (V&V) activities in the aerospace industry typically rely on time-consuming simulation-based tools. These tools can provide a probability measure for sufficiently frequent phenomena, but may fail to detect rare but critical combinations of parameters. As the complexity of modern space systems increases, this limitation plays an increasing role. In recent years, model-based worst-case analysis methods have reached maturity. Without requiring simulations, these tools can fully explore the space of all possible uncertain parameter combinations, and provide guaranteed mathematical bounds on robust stability margins and worst-case performance levels. However, they give no measure of probability and may therefore be overly conservative. Conversely, probabilistic µ-analysis combines worst-case information with probability measure. As such, it tends to bridge the analysis gap between Monte Carlo simulations and deterministic worst-case approaches. The latest developments in probabilistic µ-analysis have all been devoted to stability margins for Single-Input Single-Output (SISO) systems. This paper addresses the extension of probabilistic gain, phase, disk and delay margins to multi-variable analysis for Multi-Input Multi-Output (MIMO) control systems. To validate the proposed approach, an in-depth analysis is conducted on an academic benchmark. The analysis capability for higher-order systems is also evaluated on two more realistic satellite models.