A sampled-data control system is a practically important system, where a continuous-time plant is controlled by a discrete-time controller. In this talk, the following two problems are considered: 1. Analysis of passivity degradation under the sampling of a nonlinear system; 2. Analysis and control of a sampled-data system whose sampling interval is uncertain and variable with time. The first problem is motivated by the well-known fact: passivity is lost under discretization even if the original continuous-time system is passive. Although this property has been discussed qualitatively, a quantitative study is made here on three types of discretization. The result can be used, for example, the choice of the sampling interval. The second problem has its motivation in networked control and embedded control, where uniform sampling is not expected. In the approach to be introduced, the problem is described in a robust linear matrix inequality, which is a parameter-dependent linear matrix inequality to be satisfied for all the possible parameter values. Although the robust linear matrix inequality is intractable directly, a tractable sufficient condition can be obtained with the mean-value theorem and the S-procedure. Conservatism of the approach can be reduced to any degree. This is a joint work with Hisaya Fujioka.