This paper develops a multivariable control technique for low-level control of an intelligent hospital bed. First, multivariable hospital bed models, nominal, upper bounded and lower bounded models, are obtained via an experimental identification procedure. Based on the obtained nominal model, the triangular diagonal dominance (TDD) decoupling technique is applied to reduce a complex multivariable system into a series of scalar systems. For each scalar system, an online adaptive control strategy is then developed to cope with system uncertainties. Compared to the conventional control method, real-time experimental results showed that our proposed multivariable control technique achieved better performance. Experimental results also confirmed that desirable system performance was guaranteed under system uncertainty conditions.