Plum ( prunusdomestica. ) is a fruit from the drupe family. There are more than one thousand varieties of this fruit in the world that are different in terms of color, shape and size. Due to the high respiration intensity and high rate of maturity process, plum fruit has a short storage life, and hence must be supplied to the market in a short time after harvesting. For grading agricultural products it is needed to measure some of their properties directly or indirectly. In general, evaluation of product quality is done by destructive and nondestructive methods. However, nowadays non-destructive methods are taken into consideration because of their advantages. Among various methods for nondestructive evaluation of agricultural products, mechanical method is widely used to estimate the quality of the product. Impact is one of the nondestructive mechanical methods for determining qualitative indices of agricultural products. In this method, a dynamic input is applied to the product. This input could be exerted by free fall of the product or throwing it from a specific height onto a force sensitive surface which is called free impact method. If the dynamic input is applied by a shaped impactor in a controlled manner, it is called hammer or forced impact method. In this study, the effect of maturity period on variation of firmness and soluble solids content (SSC) of plum fruit (Khansar variety) was investigated by using an online free impact signal measurement system. The system consisted of three main parts: a conveyor unit for traorting and throwing the sample, a load cell and a data acquisition unit. The experiments were carried out at three speeds of 1, 1.5 and 2 m s -1 on 648 samples. After collecting the mechanical signals, the considered indices were measured using the reference and standard methods. The Magnes-Taylor standard destructive test was conducted for measuring the samples’ stiffness by a probe with cross-sectional diameter of 4 mm using a comprehensive tension-compression test apparatus. The maximum force for penetrating into the fruit sample was considered as the firmness. The values of SSC were measured by a refractometer from the plum extract. Preliminary iection of the signals revealed that the samples with higher stiffness values created the signals having the higher first peak values. In order to model F and SSC three different regression analysis methods including the partial least squares (PLS), the combination of principal component analysis and artificial neural network (PCA-ANN) and neuro-fuzzy were used and compared. The results showed that estimation models for F led to better results in comparison with SSC estimation models at every test speeds. Since the SSC is a chemical index, acceptable results could not be obtained for estimating that using the mechanical methods. The best result for estimating F belonged to neuro-fuzzy modeling method at the speed of 1 m s -1 with R 2 P =0.753, RMSEP=0.506 N and SDR=1.730. After neuro-fuzzy, the PCA-ANN and the PLS methods led to the best results, respectively.According to the results, the traorting and throwing velocity of 1 m s -1 is recommended for estimating the ripeness degree of plum fruit using the nondestructive free impact method. Keywords : Artificial neural network, firmness, free impact method, neuro-fuzzy, non-destructive, partial least squares, plum, soluble solids content.