Recycling of livestock manure to agricultural land may reduce the use of synthetic fertilizer and thereby enhance the sustainability of food production. However, the effects of substitution of fertilizer by manure on crop yield, nitrogen use efficiency (NUE), and emissions of ammonia (NH₃ ), nitrous oxide (N₂ O) and methane (CH₄ ) as function of soil and manure properties, experimental duration and application strategies have not been quantified systematically and convincingly yet. Here, we present a meta-analysis of these effects using results of 143 published studies in China. Results indicate that the partial substitution of synthetic fertilizers by manure significantly increased the yield by 6.6% and 3.3% for upland crop and paddy rice, respectively, but full substitution significantly decreased yields (by 9.6% and 4.1%). The response of crop yields to manure substitution varied with soil pH and experimental durations, with relatively large positive responses in acidic soils and long-term experiments. NUE increased significantly at a moderate ratio (<40%) of substitution. NH₃ emissions were significantly lower with full substitution (62%-77%), but not with partial substitution. Emissions of CH₄ from paddy rice significantly increased with substitution ratio (SR), and varied by application rates and manure types, but N₂ O emissions decreased. The SR did not significantly influence N₂ O emissions from upland soils, and a relative scarcity of data on certain manure characteristic was found to hamper identification of the mechanisms. We derived overall mean N₂ O emission factors (EF) of 0.56% and 0.17%, as well as NH₃ EFs of 11.1% and 6.5% for the manure N applied to upland and paddy soils, respectively. Our study shows that partial substitution of fertilizer by manure can increase crop yields, and decrease emissions of NH₃ and N₂ O, but depending on site-specific conditions. Manure addition to paddy rice soils is recommended only if abatement strategies for CH₄ emissions are also implemented.