A weather generator can be defined as a statistical model for the daily sequences, usually considering the multiple weather variables, such as precipitation, radiation, humidity, and temperature. The conventional weather generator has been mostly used the dry and wet days as random variables. But, the generated time series can contain a relatively large error in the dry and wet spell, due mainly to the key variable of precipitation. Fortunately, a Weather Generator Using Dry and Wet Spells (WGDWS) can be served as a new weather generator, particularly for the dry and wet spells as random variables. In this study, the usability and accuracy of WGDWS were evaluated on the daily weather data from 16 stations located in five major climate regions of China. The statistical values of generated and measured meteorological variables and the statistical values generated by WGDWS and DWSS (a weather generator based on the dry and wet daily transition probability developed by the research group) were compared. A significance test was performed on the monthly values of each climate variable, and the dry- and wet-spell lengths generated by WGDWS. The results showed that there were no significant differences from the measured values. The WGDWS has been widely expected to generate a long series of daily weather data, thereby meeting the climate, moisture, and crop physiological models. Specifically, 93.8% and 96.4% of the absolute errors in the monthly maximum and minimum temperatures were within 0.5°C of the measured values. Some 95.8% of the absolute errors were < 1 d in the monthly rainfall days. Also, 91.7% of monthly rainfall was < 10 mm, where the absolute errors in the monthly rainfall were higher for the tropical and subtropical monsoon climates, exceeding 20 mm. The absolute error in 70.3% of the total monthly solar radiation was within 2 MJ/m2, where the error was larger (> 4 MJ/m2) for the subtropical monsoon climate (Tengchong, Yunnan Province). According to the length of the wet and dry spells, the simulated values of the maximum monthly wet, average dry, and wet spell were very consistent with the measurement, with the average absolute errors < 1 d. There was a slightly larger average relative error of the maximum dry spell of the month (4.16 d). In climate type, the WGDWS was used to simulate the temperate and subtropical monsoon climates better than the temperate continental and tropical monsoon climates. A comparison was made on the monthly error distributions of the daily simulation sequences generated by the WGDWS and DWSS (a weather generator using the dry and wet daily transition probability that was previously developed by the research group). There was a consistency in the error distributions for the monthly maximum temperature, monthly minimum temperature, and monthly total solar radiation. The WGDWS simulation of monthly precipitation days was better than that of DWSS, where there was a relatively smaller error of WGDWS under the condition of equal probability. Nevertheless, the DWSS simulation of monthly precipitation was better than that of WGDWS. According to the dry- and wet-spell lengths in the daily series, there was a very close error distribution of the monthly average dry spell of the WGDWS and DWSS. The relative error of WGDWS was smaller than that of DWSS in the maximum dry and wet spell, and the average wet spell. Anyway, the WGDWS simulations were better than those of DWSS in most climate types, except for the average dry spells of the subtropical monsoon and tropical monsoon climates. The WGDWS can be widely expected to better simulate the maximum dry and wet spell, and the average wet spell, compared with the DWSS. Therefore, the WGDWS can accurately reflect the actual conditions in the long-term drought or long-term rainy weather.