Non-point source pollution (NSP) has become the bottleneck in the sustainable development of agricultural. In this study, a fuzzy-boundary interval double-side chance-constrained programming (FBIDCCP) is established for agricultural non-point source pollution control, in order to coordinated the relationship between agricultural development and environment protection. To overcome the uncertainty of randomness and fuzziness in the agricultural system, the developed method integrates the concept of fuzzy boundary intervals (FBI) into the framework of inexact double-hand-side chance-constrained programming (IDCCP). Based on probability density functions (PDFs) and fuzzy set theory, this method introduces the default risk level (pk) and the confidence level (α-cut) to transform random parameters appearing on both sides of constraints, as well as certain interval parameters with fuzzy boundaries on the upper and lower bound. The solution is achieved based on interactive algorithms and the fuzzy vertex analysis method. A case study is applied to verify the effectiveness of this method, and 36 decision scenarios is set, considering different combinations of pk and α-cut. The results indicate that, the agricultural system obtains the maximum net economic benefits under the most loose environmental emission standard (pk = 0.1) and the highest confidence level (α = 0.8). FBIDCCP extends the applicability of traditional methods, and provides effective and flexible decision support for agricultural non-point source pollution control, supporting the sustainable development of agriculture.