The objective of this work was to determine the best technique, sampling unit, and the number of samples to compose a conventional sampling plan for the cucurbit borers, Diaphania spp. (Lepidoptera: Pyralidae), and for hymenopteran parasitoids on cucumber. This research was carried out in 10 commercial cucumber crops fields from July to December 2000 in Tocantins, Minas Gerais State, Brazil (21° 11′ 15″ S; 42° 03′ 45″ W; altitude 363 m). The sampling methods studied were beating on a tray, direct counting of insects on the lower leaf surface, and whole leaf collection. Three sampling units also were studied: leaves from a branch located in the apical, median, or basal third of the canopy. The best sampling systems, which included the best technique and sampling unit, were determined based on the relative variance and the economic precision of the sampling. Once the best sampling systems were established, the numbers of samples to compose the conventional sampling plans were determined. The more suitable sampling system for the larvae of Diaphania spp. in cucumber plants was beating a leaf of the median third of the canopy on a plastic tray. One leaf must be sampled for every 50 plants in a crop. The more suitable sampling system for hymenopteran parasitoids in cucumber plants was to directly count the adults on one leaf of the median third of the canopy. One leaf must be sampled for every 74 plants in a crop.