Markov Chain Monte Carlo (MCMC) is a class of algorithms used to sample from a probability distribution. It constructs a Markov chain that has the desired distribution as its equilibrium distribution. With sufficient sampling, the state of the chain can be used to approximate the desired distribution.