The Monte Carlo sampling technique, in which polymers are sampled randomly from an infinite number of polymer molecules in the reaction system, is used to make computer simulations for the molecular-weight distribution (MWD) change during crosslinking and degradation processes. In this method, since one can investigate each polymer molecule directly, very detailed structural information can be easily obtained. This present method is not restricted by the condition of long primary chains with a low degree of crosslinking and degradation, and it is also straightforward to account for the effect of the residence time distribution. When the initial MWD is extremely narrow, say P-w/P-n <1.1, a multimodal distribution may be obtained by crosslinking reactions; however, in general, the crosslinking and/or degradation reactions tend to produce a unimodal MWD as long as the size and structural dependence of crosslinking and scission reactions is not important. When the expected crosslinking density is not the same for all primary polymer molecules, namely there exists a crosslinking density distribution among the primary polymer molecules, the average crosslinking density at the gel point is smaller than the value of 1/P-wp that is expected from the Flory-Stockmayer theory, where P-wp is the weight-average chain length of the primary polymer molecules.