The aim of this study is to clarify changes in the ecosystem carbon cycle in response to predicted global warming in various ecosystems including semi-natural grassland. To clarify responses of the whole ecosystem to warming in a semi-natural cool-temperate grassland, we conducted an in situ warming experiment and examined plant growth and CO₂ flux responses. Five pairs (control and warmed plots) of Zoysia japonica plots were established. Warmed plots were warmed using infrared heaters from June to November 2009. Once a month, aboveground biomass (AGB) of Z. japonica was estimated using the point frame method. Net ecosystem production (NEP) and ecosystem respiration (Re) were determined from CO₂ flux measured using the closed chamber method. Each month, relationships were obtained between photosynthetic photon flux density (PPFD) and NEP (PPFD-NEP curve) and soil temperature (ST) and Re (ST-Re curve). Monthly cumulative NEP, Re, and gross primary production (GPP, sum of NEP and Re) were calculated using these relationships and continuously recorded PPFD and ST data. Although there were some mechanical problems when using infrared heaters, the soil temperature in the warmed plots where infrared heaters worked well was an average of 2.3°C higher than the control plots. This suggests the heating method using infrared heaters is applicable to grassland ecosystems. AGB in the warmed plots tended to be higher (by a maximum of 70%) than in the control plots throughout the experimental period, suggesting that experimental warming affected the phenology and extended the growth period of Z. japonica. Initial slope and light compensation point of PPFD-NEP curve were significantly affected by the warming. Monthly cumulative GPP in the warmed plots tended to be higher (by a maximum of 32%) than in the control plots. This is partly explained by the increased biomass and changed photosynthetic characteristics in the warmed plots. Although not all parameters of the ST-Re curve were affected by warming, monthly cumulative Re in the warmed plots tended to be higher (by a maximum of 35%) than in the control plots. As a result, monthly cumulative NEP in the warmed plots tended to be higher than in the control plots, especially in July (approximately 70% higher) and October (approximately 100% higher). These results suggest that experimental warming affected the carbon cycle in semi-natural cool-temperate grassland by changing the phenology and photosynthetic characteristics of Z. japonica. To promote a better understanding of whole ecosystem responses to predicted warming, a longer term experiment and more detailed descriptions of carbon dynamics including the belowground part of the ecosystem are needed.