In industrial metrology, measurement and assessment of standard profile features such as planes, cylinders, cones, etc. is an important part. Many algorithms for fitting and noise-reduction of range data from single feature have been proposed. However, according to the definition of workpiece or requirement of drawing, best-fit operation of a pattern of profile features to the true profile should be taken, which requires considering several different features as a whole to fit to the true profile. Currently there’s not so much papers related with this problem. On basis of algorithms for fitting of single profile feature and noise reduction methods, an algorithm for best-fit of a pattern of profile features was presented in this paper. Since it is a nonlinear optimization problem, a method of finding initial estimates was also related in this paper, which is also helpful for the fitting of single cylinder, cone and other standard quadric surfaces.