Straw return is widely recommended as an effective measure to increase the amount of soil organic carbon (SOC) and improve soil fertility in agroecosystems. However, the temporal dynamics of the original and newly derived SOC during long-term straw return remain uncertain. This study aimed to (i) determine the temporal dynamics of original and newly derived SOC in various soil organic matter (SOM) fractions during cultivation and (ii) evaluate the effect of continuous straw return on the contents of original and newly derived SOC. These aims were achieved by analyzing topsoil samples (0–20 cm) in a 17-year field experiment using a continuous maize (Zea mays L.) cropping system. Three fertilizer treatments were tested: no fertilizer (NF), mineral fertilizer (NPK), and mineral fertilizer with straw return (NPKS). The SOC was physically separated into three density fractions: the free light fraction (fLFOC), occluded light fraction (oLFOC), and heavy fraction (HFOC), and their δ13C values were determined. The original and newly derived SOC of each fraction and the rate of conversion of exogenous C inputs were evaluated. The SOC content in the bulk soil and C fractions generally increased continually with cultivation time in the NPKS, indicating that the soil had not reached maximum C sequestration. NPKS significantly increased the C contents of fLFOC, oLFOC, and HFOC by 38.6%, 43.0%, and 11.6%, respectively, relative to the initial soil after the 17-year experiment. The newly derived SOC content was significantly higher in NPKS than in NF and NPK, and the rate of conversion of exogenous C inputs to SOC was the lowest in NPKS. The annual rate of loss of original SOC in NPKS was lowest for the light fractions (fLFOC and oLFOC) and highest for HFOC, implying that the return of straw moderated the mineralization of original SOC in the light fractions but increased the decomposition of original SOC in the heavy fraction via the priming effect (PE). The average rate of conversion of exogenous C inputs remained the highest in HFOC (23.5%), lowest in oLFOC (0.6%), and intermediate in fLFOC (4.1%) throughout the study. Thus, the stable HFOC pool retained more straw C, despite the stronger PE. Our results demonstrated that long-term straw return continuously increased the proportion of new SOC in the bulk soil and density fractions and increased the decomposition of the original SOC in the stable SOC fractions. These findings provide further detailed insights into the mechanisms of SOC stabilization in agricultural soils.