The Taiwan Cancer Registry (TCR) is a nationwide population-based registry that collects the data of patients with newly diagnosed cancer from hospitals with ≥50 beds. TCR data are high quality in terms of completeness and timeliness. However, accuracy is also a crucial quality indicator. This study evaluated the accuracy rates of selected 55 major items in the long-form TCR data between 2014 and 2016 with 700 reported cases randomly selected from 25 long-form-reporting hospitals. We calculated the accuracy rates of the reported data by employing a reabstracted chart review. Among the 55 items, the accuracy rates of 38 (69%) were at least 95%, those of 10 (18%) were between 90% and 95%, those of 5 (9%) were between 85% and 90%, and the remaining 2 (4%) were between 80% and 85%. This demonstrates a high degree of accuracy in the TCR long-form data.