在matplotlib中，日期时间轴也使用数字，即
Matplotlib使用浮点数表示日期，指定从0001-01-01 UTC起的天数加1。
许多函数如plot、scatter、bar等自动将datetime对象转换为这些数字，而许多辅助函数如axvspan直到matplotlib的最新版本才进行这种自动转换。
所以在matplotlib 3中你可以很容易地

ax.axvspan(datetime(2019,3,1), datetime(2019,3,31))

但如果使用较旧的版本，您需要手动执行，使用matplotlib.dates.date2num，例如

ax.axvspan(date2num(datetime(2019,3,1)), date2num(datetime(2019,3,31)))

一些完整的例子：

from datetime import datetime
import matplotlib.pyplot as plt
from matplotlib.dates import date2num

fig, ax = plt.subplots()

ax.plot([datetime(2019,2,14), datetime(2019,4,26)], [1,2])

ax.axvspan(date2num(datetime(2019,3,1)), date2num(datetime(2019,3,31)), 
           label="March", color="crimson", alpha=0.3)

ax.legend()
fig.autofmt_xdate()
plt.show()

感谢ImportanceOfBeingErnest的帮助，对于任何一个阅读这是我如何用冬季/夏季背景和单一标签的情节：

yearlist = ['2013','2014','2015','2016','2017','2018','2019']

fig, ax = plt.subplots(figsize=(20, 10))
for i in range(len(yearlist)):
    if yearlist[i] == '2013':
        ax.axvspan(date2num(datetime.datetime(2012,10,15)), date2num(datetime.datetime(int(yearlist[i]),5,15)), 
           label="winter (15 october - 15 may)", color="crimson", alpha=0.3)
        ax.axvspan(date2num(datetime.datetime(int(yearlist[i]),5,15)), date2num(datetime.datetime(int(yearlist[i]),10,15)), 
           label="summer (15 may - 15 october)", color="blue", alpha=0.3)
    else:
        ax.axvspan(date2num(datetime.datetime(int(yearlist[i-1]),10,15)), date2num(datetime.datetime(int(yearlist[i]),5,15))
           , color="crimson", alpha=0.3)
        ax.axvspan(date2num(datetime.datetime(int(yearlist[i]),5,15)), date2num(datetime.datetime(int(yearlist[i]),10,15)), 
            color="blue", alpha=0.3)