Abstract: Scientifically and accurately depicting the development trend and spatial characteristics of China’s local government debt scale is an important premise for scientifically formulating policies and measures to prevent and resolve local governments’ debt risks. Based on the official data of local government debt and GDP at city level, using the compiled data in 2016—2020 of local government debt ratio of 355 cities, this paper analyzed the temporal and spatial evolution characteristics of local government debt ratio in China by using nuclear density estimation, Dagum Gini coefficient, standard deviation ellipse, exploratory spatial data analysis, coefficient of variation and spatial panel model. The findings are as follows: First, the debt ratios of local governments in the whole country and the four regions—East, Central, West and Northeast China—all showed a rapid upward trend, and the debt ratios and growth rates of local governments in each region showed a spatial pattern of “Northeast > West > East > Central China”, and the whole country and all regions showed gradient effects in different degrees, and there were multipolar differentiation phenomena. Second, during the sample period, the overall difference, intra-regional difference and inter-regional difference of the debt ratio of local governments showed a steady downward trend, among which the main contributor to the overall difference was inter-regional difference, followed by super-variable density, which also had a greater impact, and finally intra-regional difference, which had a smaller impact. Third, the location distribution of local government debt ratio showed a spatial pattern of northeast-southwest, with the distribution center at the junction of Henan and Hubei provinces. Fourthly, the spatial agglomeration mode of local government debt ratio is dominated by “high-high” agglomeration and “low-low” agglomeration, among which “high-high” agglomeration cities are relatively few and mainly distributed in northeast, north, southwest and northwest regions. Fifth, the debt ratio of local governments in the whole country and four regions not only has a σ convergence trend, but also has the characteristics of absolute β convergence and conditional β convergence, and the spatial spillover effect plays an important role in β convergence, but this effect is significantly different in different regions.