Sampling and testing for microorganisms in foods is a risk management strategy used to evaluate whether a food safety system achieves the appropriate level of control. Sampling plans can be derived to meet the consumer's and/or producer's quality requirements, and have been traditionally designed using classical acceptance sampling theory. This chapter describes in detail and illustrates the methodologies to derive the different type of sampling plans used for microbiological criteria in foods; namely, the two-class (based on prevalence, on concentrations, and with an enrichment step) and three-class attributes sampling plans and the variables sampling plans. The validity of the assumption that the log microbial concentration is normally distributed among food units with a variance that is approximately stable batch to batch is questioned; and, within this context, new modelling trends based on more realistic assumptions are examined that consider the clustering of bacteria, the intrinsic variability among food batches and the use of past monitoring microbial data.