In order to obtain the microphysical features of snowflakes under the natural conditions by using the self-developed sensor, precipitation microphysical characteristics, capture probability, and space distribution retrieve algorithm based on the velocity of snowflakes are discussed; and the shape, size, velocity, axial ratio, orientation and the space distribution of snowflakes are investigated by the field observation. Results show that the snowlakes may take needles, flakes, and oblate ellipsoids shapes; the falling velocity of snowflakes will increase slightly with inereasing diameter, their axial ratio may increase linearly with the diameter, and their orientations are easily influenced by the wind; the canting angle may be 0.9° and tends to be nearly symmetric with a standard deviation around 13.2°. Empirical equations of velocity and diameter, and of axial ratio and diameter are fitted by nonlinear fitting and linear fitting of observed data respectively. Compared with previous model proposed by authors abroad, this paper further provides the microphysical characteristics of snowflakes with localization features. The above conclusions can be of significance for the research on cloud precipitation physics, ground validation of weather radar, assessment of precipitation-induced microwave attenuation, etc.