Globally, fishways perform better when the target fish species are salmonids rather than non‐salmonids. In the past few decades, engineers and dam managers have attempted to identify an ideal fishway type that can fit all migratory orders. To determine a general rule for selecting fishway types, we classified all fish orders into two groups (salmonids and non‐salmonids) and analyzed the effect of different fishway types (i.e., nature‐like [NL], vertical slot [VS], pool and weir [PW], and denil [DL]) on upstream migration for both groups by conducting a meta‐analysis based on data from 64 studies of 76 fishways. The fishway performance was presented in terms of attraction efficiency (AE) and passage efficiency (PE). The results of a random‐effects model, including 257 and 299 estimates of AE and PE, respectively, showed fishway types only had a significant effect on AE for non‐salmonids. Technical fishways had a higher AE than NL fishways for salmonids, but were not applicable to non‐salmonids. The passage results of four fishway types indicated that the PW fishways exhibited the best performance for salmonids, but the poorest performance for non‐salmonids. Only DL fishways had a PE of over 50% for non‐salmonids, whereas all fishway types had that of over 60% for salmonids. Variations in the slope and length of VS fishways could significantly affect PE for both groups, however, the phenomenon for salmonids disappeared when merging all fishway types. Elevation changes had a significant effect only on the condition that salmonids passed through VS fishways. The results of this study play an important role in designing appropriate fishway types for different downstream fish fauna and can be better generalized by complementing quantitative evaluation studies of various fishways as much as possible.