Evaluation of arable land ecosystem services capacity and arable land-use intensity (ALUI) is important for recognising key regional factors that impact arable land attributes changes, which is crucial for planning sustainable patterns of arable land use. The chronic lack of coordination between these two types of evaluation studies has made it hard to provide enough information for developing arable land-use management and control policies. Here, we generated a 1-km-grid map of arable land potential yield and county-level arable land productive-capacity. The impact of land-use and land-cover change on county-level total arable land productive-capacity during 1990–2010 had been estimated. Then we determined the aggregation–distribution characteristics of four indexes (i.e. average arable land potential yield, average ALUI, total arable land area and arable land productive-capacity reserves) at the county-level by the k-means algorithm to assess the regional coordination between arable land productive-capacity protection and arable land use. The results show that during 1990–2010, land-use change led to arable land productive-capacity decreases in 2007 of China’s 2733 counties (nearly 73.5% of the total counties’ count). Most of these counties are in central and southern China, and their corresponding arable land productive-capacity decrement rates are generally < 6.15%. Counties with decrement rates > 6.15% are mainly in the Yangtze and Pearl River delta regions. The geographical detector shows that county-level arable land-area change is a primary factor that drives county-level arable land productive-capacity increase. Its determinant power can be quantified as 74.154%. In contrast, its determinant power to county-level arable land productive-capacity decrease is only 38.542%, which demonstrates that occupy high-capacity arable land and supplement low-capacity arable land have a greater role in causing reduction of county arable land productive-capacity. Total arable land productive-capacity and use intensity show only slight determinant power to county-level arable land productive-capacity decrease. It indicates that insufficient attention has been paid to the protection of arable land productive-capacity and the farmers’ willingness in the implementation of China requisition-compensation balance policy. The aggregation–distribution characteristics show that arable land productive-capacity protection is in coordination with arable land use for most of the provinces in China. Arable land with high or medium potential yield tends to be used at high or medium intensity. A lack of coordination is most evident in the insufficient ALUI, particularly in seven of the thirteen major grain-producing provinces. Other evidence of weak coordination is in the low potential yield versus high farming conditions and willingness, where unsuited planting modes should be prevented. Lastly, challenges for exploring sustainable arable land use path have been discussed. This study is greatly instructive for recognising interrelations between natural conditions and arable land-use patterns and for exploring shortcomings that impede regionally sustainable arable land use. •Impact of LUCC on county-level arable land productivity during 1990–2010 is estimated.•Difference in determinant power to arable land productivity decrease and increase is detected.•Regional coordination between arable land productivity and use intensity is assessed.•Correlation between arable land potential yield and its distance to urban is assessed.•Generate county-level normalised total arable land productive capacity dataset.