Despite numerous studies on methanotrophs, virtually nothing is known about their spatial heterogeneity in nature. These patterns, however, have strong influences on the interpretations made from analysing microbial processes and community structure. Here we report the first use of geostatistics to analyse the spatial heterogeneity of methanotrophs in a rice field soil (Vercelli, Italy). We used the gene encoding the particulate methane monooxygenase, pmoA, for terminal restriction fragment length polymorphism (T-RFLP) analysis. The profiles obtained were compared using a pseudo-variogram analysis to study autocorrelation as a function of distance. We demonstrated that there was no large-scale spatial structure at this study site, but a micro-scale spatial structure could not be excluded. A species accumulation curve with all terminal restriction fragments revealed that even 75 samples were insufficient to cover the diversity of methanotrophs in a rice field. However, a species accumulation curve of methanotrophs defined as operational taxonomic units validated from a clone library with 90% coverage demonstrated saturation after approximately 15 samples. The results of this study have consequences for studying the diversity and function of methanotrophs. In this agroecosystem population structure showed no spatial pattern implying that both a systematic and random sampling design would be adequate.