@yangguoshe wrote:
Thank your replaying, I want to know if we can deduce the basic function (De Boor’s Algorithm)of each sub segment of the spline curve(AB BC CD DE EF ), and rely on the basic function to deduce the new control points, and to realize the complete coincidence between the new two segments of the spline curve and the original one

Recreating spline segments between Fit Points is a straightforward process. A Fit Point is a knot location so the segment between two consecutive fit points can be recreated as a basic Bezier style curve. And this is true regardless of the Knot Parameterization method.
The tangent direction (acquired via Xline in the screencast) could be retrieved with VisualLisp’s vlax-curve-getFirstDeriv function. The Fit Point location, and the two intersection with the related Control Vertices frame can recreate the segment with perfect fidelity.
Breaking a spline mid span becomes more involved – where the math and/or geometry manipulation increases in complexity. Still doable, if someone were so inclined, but tough to justify with the tools already available – either as a Command call or via the Managed ARX API.
Perhaps one of the programmers versed in both .NET and ALisp could encapsulate the Curve.GetSplitCurves Method.
public virtual DBObjectCollection GetSplitCurves(Point3dCollection point);
************************************************************
May your cursor always snap to the location intended.