Since its discovery in the southeastern United States, the invasive plataspid Megacopta cribraria (F.) (Hemiptera: Plataspidae) has infested soybean (Glycine max (L.) Merrill) fields in often very high numbers. To optimize sampling plans, sweep-net and beat-cloth sampling was conducted in soybean fields in South Carolina during 2012 and 2013. Across all fields, densities averaged 7.2 ± 0.5 (SEM) adults and 4.5 ± 0.4 nymphs per 20 sweeps and 5.5 ± 0.3 adults and 4.5 ± 0.3 nymphs per 1.83 m of row. Coefficients of Taylor's power law were used to generate sampling plans for population estimates and sequential sampling plans for pest management decision making. At an economic threshold of one nymph per sweep, optimum sample sizes were 184, 48, and 22 within 10, 20, and 30% of the mean with the sweep-net method. At the corresponding threshold for the beat cloth (24.7 nymphs per 1.83 m of row), optimum sample sizes were 239, 62, and 29 within 10, 20 and 30% of the mean, respectively. At all adult and nymph densities, fewer sweep-net samples were required for population estimations compared with the number of beat-cloth samples. Sequential sampling reduced the sample size required to reach a management decision for the sweep net and beat cloth compared with a fixed sampling plan. The sweep-net method was more cost reliable for population estimation at low densities of both life stages, with the beat cloth becoming more cost reliable as populations increased. The beat-cloth method was more cost reliable than sweep-net sampling across all densities and life stages for pest management practices. These results may be used by researchers, county Extension agents, consultants, and farm managers to both facilitate sampling and improve reliability of M. cribraria estimates for research purposes and for pest management.