Abstract Soil respiration (Rs) during daytime (Rsday) and nighttime (Rsnight) remains poorly understood owing to the limited availability of high‐resolution temporal measurements of Rs. Using continuous automatic chamber measurements at a half‐hourly interval, we explored Rs and its temperature sensitivity (Q10) during the daytime and nighttime over 10 years and determined the best sampling timeslot in a temperate deciduous broadleaf forest. We demonstrate a higher Rsnight than Rsday in spring and summer, a lower Rsnight than Rsday in autumn, but an insignificant difference between them in winter, contrasting with the consensus of a higher Rsday than Rsnight. Such discrepancy may result from the effect of factors other than temperature (e.g. reallocation of photosynthesis carbon, precipitation and soil moisture). We also reveal significant differences in apparent Q10 between seasons, indicating the highest Q10 in spring and significantly higher Q10 at night than at day in autumn. The Rs measured around midday (9:00–11:00 for autumn and 12:00–15:00 for other seasons) and midnight (20:00–22:00 for autumn and 0:00–2:00 for other seasons) provides the lowest bias to the daily mean Rs, whereas the measurement timeslots around midday are preferred due to practical considerations. Our findings highlight the significant diel variation of Rs and its temperature sensitivity across seasons and the necessity to account for Rsnight for a more accurate estimate of soil carbon budget. Read the free Plain Language Summary for this article on the Journal blog.