The ionospheric storm has been studied for over 50 years and different physical mechanisms have been carried out to explain the different responses of ionosphere. Some studies focused on the ionospheric disturbance in a specified latitude zone since the ionospheric storm shows significant dependence on latitude. Using TEC data from Madrigal Database in the European sector during 156 geomagnetic storms between 2001 and 2010, the statistical features of the ionospheric storms over 5 latitude zones, from equator to auroral oval, are presented in this paper. SYM-H data is used to determine all the geomagnetic storms between 2001 and 2010. Altogether 156 storms were identified, whose main phase undergo a single step growth according to the criterion of Kamide et al (1998). Diurnal TEC data over 5 latitude zones are required through spatial-average and time-average process in both storm and peacetime. The difference of diurnal TEC in storm and average diurnal TEC in peacetime is identified as the response of ionosphere. If the difference is more (less) than 25 Vo of the average diurnal TEC in peacetime for more than 3 hours, the storms are classified as positive (negative) storms or P (N) storms. Some ionospheric storms show initial positive storms followed by negative storms, which are classified as PN storms. A small number of storms, which are found to have initial negative storms followed by positive storms, are classified as NP storms. NS storms are identified as non-significant ionospheric response during geomagnetic storm. From the statistical features of ionospheric storms over 5 latitude zones, including the dependences of ionospheric storms on latitude, season, phase of geomagnetic storms and local time, and time dependence of the onset of ionospheric storms, the results show that: 1) the ionospheric storms show significant features in latitudinal distribution, the ratio of positive storms to negative storms decreases with increasing latitude with one exception that the ratio in equatorial area is even lower to that in higher latitudes; 2) the number of positive storms decreases in recovery phase when compared with main phase, and the number of negative storms shows opposite pattern. However, the number of positive storms increases in equatorial area during recovery phase; 3) the MPO of positive storms in low and middle latitude mainly appears in daytime and nighttime MPO tend to cause negative storms; 4) negative storms tend to occur in night sector, a significant feature of the negative storms at mid-high latitudes is the so-called 'forbidden time interval', which shows latitudinal difference in the local time distribution. And positive storms shows more scattered local time distribution. Some of the results in this paper are compared to former studies and given explanations; however some results still need further study. This research contributes more understanding of the features and mechanisms of ionospheric storms in different latitudes, and provides useful information for modeling and forecasting of ionospheric disturbance.