Effective predation of coccinellids species is essential for plant protection programs in the field or glasshouse. The purpose of this study was to estimate deviations of predatory effect from predicted values based on the predatory effect model proposed by previous researchers. Two coccinellids species were used to cope with six different pests. There were differences between realized and theoretical data. Especially for aphid species there were very low deviations and in one case there were a positive deviation for both predators, indicating that Coccinellidae were underestimated in previous studies. As a conclusion, prediction model must be used as a guide for practical purposes and not for theoretical modeling. The choice of a predator species must be a result of extensive testing on the prey because differentiation of predatory effect is common to coccinellids and depends also on the locality of a species. Coccinella septempunctata found to be a better predator. The model predictions were in accordance only for M. persicae and M. euphorbiae measurements and predators-prey relationships were different on different plant leaves.