The P efficiency, crop yield, and response of wheat to arbuscular mycorrhizal fungus (AMF) Glomus caledonium were tested in an experimental field with long-term (19 years) fertilizer management. The experiment included five fertilizer treatments: organic amendment (OA), half organic amendment plus half mineral fertilizer (1/2 OM), mineral fertilizer NPK, mineral fertilizer NK, and the control (without fertilization). AMF inoculation responsiveness (MIR) of wheat plants at acquiring P were estimated by comparing plants grown in unsterilized soil inoculated with G. caledonium and in untreated soil containing indigenous AMF. Without AMF inoculation, higher crop yields but lower colonization rates were observed in the NPK and two OA-inputted treatments, and NPK had significantly (P < 0.05) lower impacts on organic C and available P in soils and thereby P acquisition of wheat plants compared with OA and 1/2 OM. G. caledonium inoculation significantly (P < 0.05) increased colonization rates with the NPK and two P-deficient treatments but significantly (P < 0.05) increased vegetative biomass, crop yield, and P acquisition of wheat as well as soil alkaline phosphatase (ALP) activity, only with the NPK treatment. This gave an MIR of ca. 45% on total P acquisition of wheat plants. There were no other remarkable MIRs. It suggested that the MIR is determined by soil available P status, and rational combination of AMF with chemical NPK fertilizer can compensate for organic amendments by improving P-acquisition efficiency in arable soils.