Several methods have been published to estimate per capita predation rates from molecular gut content analysis relying on intuitive understanding of predation, but none have been formally derived. We provide a theoretical framework for estimating predation rates to identify an accurate method and lay bare its assumptions. Per capita predation can be estimated by multiplying the prey decay rate and the prey quantity in the predators. This assumes that variation in per capita predation rate is approximately normally distributed, prey decay occurs exponentially, and predation is in steady state. We described several ways to estimate steady state predation, including using only qualitative presence‐absence data to estimate the decay rate and in addition, we provided a method for estimating per capita predation rate when predation is not in steady state. We used previously published data on aphid consumption by a ladybird beetle in a feeding trial to calculate the predation rate and compare published methods with this theoretically derived method. The estimated predation rate (3.29 ± 0.27 aphids/h) using our derived method was not significantly different from the actual predation rate, 3.11 aphids/h. In contrast, previously published methods were less accurate, underestimating the predation rate (0.33 ± 0.02 to 1.66 ± 0.8 aphids/h) or overestimating it (3.64 ± 0.30 aphids/h). In summary, we provide methods to estimate predation rates even when variation in predation rates is not exactly normally distributed and not in steady state and demonstrate that the prey decay rate, and not the prey detection period, is required.