I have EPIC array methylation data and would like to check for differential methylation.

+------+-----+
| time | id  |
+------+-----+
| t0   | P1  |
+------+-----+
| t1   | P1  |
+------+-----+
| t0   | P2  |
+------+-----+
| t1   | P2  |
+------+-----+
| ...  | ... |
+------+-----+

So far I have always used the limma package in R here, but I have problems defining the study design for the repeated measures.

group <- factor(targets$time,levels=c("t0", "t1"))
id <- factor(targets$id)
design <- model.matrix(~id + group)

Can someone help me define the design correctly or point me in the right direction if the approach is completely wrong?

Looks good to me. The last coefficient in design should represent the log-fold change (or for methylation data, the difference between M-values) between t0 and t1, which is presumably the contrast of interest here.